P2P Application Help

Transcription

P2P Application Help Sections
1
2
3
4
5
Welcome and System Requirements
Getting Started / Usage
Helper Applications/tools
Conferencing and Firewalls
Advanced Options
Notes
Welcome !
P2P (Person to person) is Virtual Lab Toolkit software which aims at facilitating communication on a person to
person level. The following types of communication are supported via external helper applications:
Audio Connection: Low speed (less than 15 kbps) audio connections. The audio component is archived using the
gsm codec of the RAT audio tool.
Video Connection: With a video camera, Low speed (less than 30 kbps) good quality video images maybe sent
with a delay of 1 or 2 seconds. Video is managed by the VIC video tool.
Scientific Talk: A web based chat system with in built support for Tex/Latex and hence mathematical symbols.
The chat server is driven by the Scientific Talk software while the client part consists of a simple Web browser!
Others are White Board, Shared Text Editor, Text Chat.
The following "optional" tools may be started and stopped from the P2P application but the user is required to
complete the connection from the client tool
Messenger Client, Desktop Sharing, H323 Application (NetMeeting), Application Sharing and MBONE SDR.
System Requirements
Hardware:
Pentium Class Computer
−
64 MB RAM
−
16 bit or better display ( over 65535 colours)
−
Working Audio (sound) Card
−
HeadSet (Earpiece and Microphone)
−
Video Capture Card or WebCam
Operating System(s):
Windows: Recommended − 9x, ME or 2000)
Linux: (Recommended − RedHat 6.x and 7.x)
Disk Space:
10 MB for minimum installation
20 MB for full installation
The P2P application and helper tools were tested in the environments specified above. However, they may work
in other environments or operating systems. See Notes.
1.0 Getting Started / Usage
1.2 Managing the AddressBook
1.3 Making a Call
1.4 Receiving a Call
1.5 Call Management Mode
1.6 Ending a Call
1.7 Unmanaging a Call
1.8 Exporting Addresses
1.9 Importing Addresses
The first time you run the P2P application, you will be presented with the following dialog box (figure 1.1).
Figure 1.1 Preferences Window
This allows you to set the following important information
1. Your nick name: Most likely this will be your first name.
2. Your full name: Your family name and other names.
3. Your e−mail address: A valid and working e−mail address. This field is reserved for future use.
4. Your preferred Web browser: On Windows this defaults to the system default browser while in Unix, it
defaults to the netscape web browser.
You can always return to this window at any time by Clicking on File −> Preferences.
1.1 Main Window
Figure 1.2 Main Window
The P2P tool "Main Window" (see figure 1.2) consists of a simple address book display that shows a list of stored
contacts. Each contact has 3 items.
1. A nick name: A short reference name.
2. The IP Address: The INTERNET hostname or IP address of the computer used by the contact.
3. The Full Name: of this contact!
The list of contacts are stored in the file p2pabook in the home directory of the user (on Windows, this defaults to
c:\).
1.2 Managing the Addressbook
The "Main Window" has two buttons for managing the "Address Book". They are "Add Address" and "Delete
Address"
Adding New Addresses are archived by clicking on the "Add Address Button": This opens a window as shown in
figure 1.3
Figure 1.3: New Address Book Entry
Deleting Addresses: First highlight the contact to delete by clicking once on it in the "Main Window", then click
on the "Delete Address" button to delete it.
Note: Deletions are immediate, without confirmation and not reversible.
Changing Entries: First highlight the contact to edit by clicking once on it in the "Main Window", then click on
the "Edit Menu" item and select "Modify".
The Addressbook window now shows the information which you can change and then save using the "Save"
button.
Note that multiple IP addresses or Hostnames are not supported. This can be circumvented by creating multiple
entries in the addressbook.
After filling in the entries, click on the "Save" button to update the addressbook file and the "Main Window"
display.
To exit the addressbook dialog box, click on the "Cancel" button.
1.3 Making a Call
From the Addressbook: Double click on the entry to open the "Invite" dialog box with the entry filled in (see
figure 1.4).
Manual Calls: Click on the "Invite" button to open the "Invite" dialog box (see figure 1.4).
Figure 1.4: Invite dialog box
You can change the displayed nick name and IP address entries.
Then select the "Call" components by clicking the checkboxes in front of the texts (Audio, Video and/or Scientific
Talk).
Entries are grayed out if the helper application is not configured (see advanced configuration). Finally click the
"Call" button to initiate the call setup process.
The P2P tool will display figure 1.5 and then try to contact the remote P2P client.
Any error involving the network connection will be shown in a dialog box.
Figure 1.5 Connecting Dialog Box
Figure 1.5 would be removed automatically when the remote P2P responds.
If the response is negative, then figure 1.6 is shown.
Figure 1.6: Call not accepted by contact
If the response is positive then the helper application(s) will be started to connect to the remote site and the
interface will change to show the "End Call" button instead of the "Invite" button.
1.4 Receiving a Call
In order to receive calls via the P2P client tool. You should always have the P2P client running. When the P2P
tool detects an incoming call. Figure 1.7 is displayed.
Figure 1.7 Incoming Call
It shows the "Full Name" of the caller along with his nickname, IP address of his computer and the various
components requested by the caller.
The call components will list the requested tools. Example: Audio, Video and Scientific Talk.
You can choose to "Accept" or "Decline" the call using the buttons at the bottom of the page.
If you accept the call, the P2P tool will automatically startup the various applications and move into the "Call
Management Mode". If you decline the caller the window show in figure 1.6 will appear.
1.5 Call Management Mode
After a call has been successfully accepted by the receiver, the P2P tool will switch into the call management
mode by automatically starting up the component tools: rat for audio, vic for video and a browser window for
scientific talk.
The "Invite" button on the "Main Window" is replaced by a "End Call" button (see figure 1.8).
Figure 1.8 Call Management Mode
The P2P tool can only startup (with suitable command line options in most cases) or shutdown the component
tools. The user is required to futher interact with the tools in order to perform tasks such as changing volume or
video size or transmit video and/or audio!
1.6 Ending a Call
When in the "Call Management" mode the current call may be ended by clicking on the "End Call" button in the
"Main Window" (see figure 1.8). The P2P tool will then stop (close down) the helper application(s) that were
started as part of the call.
The other alternative is to individually close the various tools.
1.7 Unmanaging a Call
The "File" −> "Unmanage Call" menu item permits the user to leave call management mode without ending the
call or terminating the applications in use in the call. To activate click on "File" and then "Unmanage Call". The
P2P tool will then exit "Call Management" mode without terminating applications used in the current call.
1.8 Exporting Addresses
Addresses may be exported from the address book using the "File" −> "Export" menu item. A file dialog is
presented and the user is required to fill in a file name. The contents of the address book will be written to this file
using the Comma Separated Value (CSV) text database structure.
The first line of the resulting file is a header describing the available fields.
1.9 Importing Addresses
Existing addresses may be imported into the P2P address book using the "File" −> "Import" menu item. A file
dialog is presented and the user is required to select an existing file. P2P is only capable of importing records in
the comma or semi−colon or TAB delimited text database format. The first line of the file is ignored and assumed
to be a header line.
Example:
nickname,hostname,Full name
nickname;hostname;Full name
nickname
hostname
Full Name
2.0 Helper Applications/tools
2.1
2.2
2.3
2.4
2.5
2.6
2.7
2.8
2.9
2.10
2.11
Audio
Video
Scientific Talk
WhiteBoard
Network text Editor
Text Chat
Messenger Client
Desktop Sharing
H.323 Application
Application Sharing
MBONE SDR
2.1 Audio
Audio is handled via the RAT audio tool. The following rat command line options are used:
−C "r_nick" (sets the conference name/title to the nickname of the remote user)
−N "m_nick" (sets the session name to the nickname of the local user)
−f gsm (sets the gsm codec)
r_ip/r_audioport (IP address of remote machine and destination port).
For more information on using the RAT please click here.
2.2 Video
Video is handled by the VIC video tool. The following vic command line options are used:
−B 32 (sets the maximum transmit speed to 32kbs)
r_ip/r_videoport (IP address of remote machine and destination port).
For more information on using VIC click here.
2.3 Scientific Talk
Scientific Talk is a Web based chat tool with special support for discussing mathematics. The scientific talk server
understands and can parse some latex syntax, so it can display special symbols and equations.
The scientific talk client is a standard web browser. The P2P tool will start up a Web browser with the URL of a
scientific talk server as the only argument.
For more information on Scientific Talk click here.
2.4 White Board
The WhiteBoard tool provides a canvas (board) for shared drawing or showing diagrams. The WhiteBoard
application is called WDB and uses the following command line options:
For more information on using WDB click here.
2.5 Network Text Editor
The Network Text Editor permits the joint editing of a text file. This function is provided using the NTE
application with the following command line syntax.
For more information on using NTE click here.
2.6 Text Chat
The text chat tool permits real time text conversation. This functionality is provided by the UNIX talk application
(or wintalk on Windows).
For more help on the talk application(s) click here.
2.7 Messenger Client
The p2p tool can optionally start up an instant message client tool. The various options include the Yahoo
Messenger, the Microsoft Network (MSN) Messenger, Netscape/AOL Instant Messenger (AIM) and Jabber
Client.
For more information the Jabber client click here.
2.8 Desktop Sharing
Desktop sharing is provided via the Virtual Network Computing (VNC) client/server. The P2P tool only starts up
a client to connect to an already running VNC server. The user is responsible for starting the VNC server
independent from the P2P tool.
For more information on the VNC tool click here.
2.9 H.323 Application
H.323 is the ITU standard for audio/video conferencing over TCP/IP networks. The most popular H.323
application is the Microsoft NetMeeting for Windows, while there exist other alternatives such as gnomemeeting
(based on openh323) for Linux, SunForum for Solaris.
For more information on H.323 Application click here.
2.10 Application Sharing (T.120) client
T.120 is the ITU standard for Application sharing. It is usually negotiated as part of a H.323 connection. T.120
application sharing is available in Microsoft's NetMeeting and the Solaris SunForum H.323 clients. It is not yet
avialable in the gnomemeeting/openh323 client.
For more information on T.120 application sharing click here.
2.11 MBONE SDR
The Multicast Backbone (MBONE) is used mainly in academic circles for holding multi−user meetings and
events.
For more information on using the MBONE SDR tool click here.
3.0 Conferencing and Firewalls
The use of INTERNET firewalls and/or port filtering routers is now a common security practice on computer
networks.
One undesirable feature of firewalls and port filtering routers is that conferencing or collaboration tools hardly
work across a firewall. In order to allow the use of p2p and it's helper applications across the INTERNET the
following recommendations should be implemented:
Application
P2P
RAT
VIC
WBD
NTE
talk
VNC
H323 Client
T.120 Client
Port(s)
TCP 7000
UDP 7002 and 7001
UDP 7004 and 7003
UDP 7006 and 7005
UDP 7008 and 7007
UDP 517
TCP 5900 and/or 5901
Too many
Notes
Port for P2P call setup/management
7002 for data, 7001 for control information
7004 for data and 7003 for control information
UNIX Talk
For Desktop sharing, actually 5900 + x where x is the port number
probably better to setup a H323 proxy/gatekeeper
MBONE SDR Multicast Network
Probably better to setup multicast Routing.
Table 3.1 TCP/IP ports used
For IP filtering firewalls/NAT firewalls it is necessary to pass and accept the connection port listed in table 1.0.
For masquerading firewalls the out going connections are masqueraded and for to the remote machine appear
to be coming from the firewall server, however the return path (the ports on the firewall) must be mapped back to
the internal machine.
4.0 Advanced Options
4.1
4.2
4.3
4.4
4.5
Advanced Preferences
Macro Substitution
Restoring the default Preferences
Recommended Helper Applications/tools
Files
4.1 Advanced Preferences
The Advanced Preferences dialog shown in figure 1.9 gives the user access to all aspects of the P2P tool. The
various applications and ports used may be configured.
Figure 4.1 Advanced Preferences
The full path name of applications may be specified. The double quote marks can be used to group options that
contain spaces. Example: "C:\Program Files\NetMeeting\conf.exe"
If the full path is not specified then the system program search path is checked for the application at execution
time. Port values should only have numerical values. The "P2P Communication Port" defines the TCP/IP port
value that the P2P tool will listen to and receive incoming requests on.
4.2 Macro Substitution
The P2P application has several special macros defined and the final values are substituted at program execution
time. This permits passing values to the helper applications at execution time. Macros consist of two parts (prefix
and suffix) separated by an underline "_" character, as explained below
Prefix
Definition
m
Local user/configuration value. See table 4.2.
r
Value specified by the remote user. The only values available are the nickname (nick) and port values of
requested applications to be used in the call. The values of IP address (IP) and full name (full) are those
from the local address book
remote Similar to the r prefix above.
Automatically select. Uses the local value, if the local user is the call, initiator otherwises use the remote
a
value
Table 4.1 Prefix of Macros
Macro surfix
nick
full
audioport
videoport
wbport
ntport
vncport
sturl
ip
Definition
Nick name
Full name of user
Port value for audio connection
Port value for video
port value for WhiteBoard connection
Port value of Network Text Editor
Port value of VNC server
Scientific Talk Server URL
IP address
Table 4.2 Common Macro surfix
Complete Macro
Meaning/Definition
m_nick
Nickname of local user
r_nick
Nickname of remote user
Nickname of the local user if the local user initaited the call, otherwise nickname of the
a_nick
remote user
Table 4.3 Example of complete Macros
4.3 Restoring the default Preferences
The initial values of the preferences may be restored at any time using the "File" −> "Default Preferences" menu
item. The user is required to then use the "Advanced Preferences" dialog to verify and save the restored options.
4.4 Recommended helper applications
Table 4.4 lists some possible helper applications and command line syntaxes for the various tools. Only use
entries in this table, if you have the application installed to it's default location. A version of all applications
shown in Table 4.4 are available on the CD−ROM.
Application/Tool
AUDIO
Windows
rat −C "r_nick" −N m_nick −f gsm
r_ip/m_audioport
VIDEO
vic −C "r_nick" −N m_nick r_ip/m_videoport
WhiteBoard
wbd −C "r_nick" −N m_nick r_ip/m_wbport
Shared Editor
nte −C "r_nick" −N m_nick r_ip/m_ntport
Scientific Talk
Server URL
Web browser
Text Chat
Shared Desktop
Linux
rat −C "r_nick" −N m_nick −f gsm
r_ip/m_audioport
vic −C "r_nick" −N m_nick
r_ip/m_videoport
wbd −C "r_nick" −N m_nick
r_ip/m_wbport
nte −C "r_nick" −N m_nick
r_ip/m_ntport
http://www.collaborium.org/cgi−bin/sci−talk/sci−talk.cgi
rundll32 url.dll, FileProtocol Handler m_sturl
\wintalk\wtalk32.exe r_nick@r_ip
"\Program Files\ORL\VNC\vncviewer.exe"
−shared a_ip:a_vncport
"\Program Files\WinJab\WinJab.exe"
sdr
"\Program Files\NetMeeting\conf.exe"
Messenger Client
MBONE SDR
NetMeeting Client
Application sharing
"\Program Files\NetMeeting\conf.exe"
Client
netscape m_sturl
xterm −e talk r_nick@r_ip
vncviewer −shared a_ip:a_vncport
gabber
sdr
gnomemeeting
Table 4.4 Recommended Helper Applications
4.5 File structure
User files: The user files p2pcfg (for configuration data) and p2pabook (the address book) are usually kept in the
user's home directory. For Windows, this could be c:\ (Root directory of installed disk) and for Linux it is the
user's home directory.
Application and help files: P2P requires the application to be installed in a directory with the help files available
in a sub−directory of this directory. For windows this directory is usually "C:\Program Files\P2P" and help folder
is "C:\Program Files\P2P\help" while in Linux it could be "/usr/local/p2p" with the help directory being
"/usr/local/p2p/help".
5.0 Notes
RAT version 4.x may not work very well with RAT version 3.x. Audio quality is better when both parties have
the save version of the audio tool.
The p2pport of the remote user maybe specified in the addressbook entries for hostname or IP Address by using
the syntax "hostname:p2pport". Example: server.somewhere.com:7000.
The P2P tool is a Tcl/Tk application. The source code is available on the CD−ROM. It require the Itcl, Itk and
Iwidgets Tcl extensions.
Compiling any of the helper tools from source code requires a working C compiler. The GNU C compiler (gcc) is
recommended.
Jabber is a new Instant Messenging protocol. A jabber user is capable of communicating (via the Jabber Server) to
other Instant messaging networks such as Microsoft, Yahoo and AOL Instant Messenger.
Windows
Online help for the WinJabber application is available from WinJabber HOWTO.
A PDF copy of the above documentation is available on the CD−ROM as winjab.pdf.
Linux
On−line help for the Gabber client is available from Gabber Manual.
A PDF copy of the above documentation is available on the CD−ROM as gabber.pdf.
H.323 Client or NetMeeting Tool
Windows
The NetMeeting Client has it's own built in Help.
On−line help for NetMeeting is available from Microsoft NetMeeting.
Linux
On−line information on GnomeMeeting may be obtained from GnomeMeeting.
The CD−ROM has a pdf copy of the GnomeMeeting FAQ as faq.pdf.
T.120 Application Sharing Tool
Windows
The NetMeeting Client provides application sharing facilities and has it's own built in help.
On−line help for NetMeeting is available from Microsoft NetMeeting.
Linux
Not yet available.
NTE Shared Network Editor Tool
On−line documentation for NTE is available from Network Text Editor.
The CD−ROM has a pdf copy of the complete NTE userguide as userguide.pdf.
RAT Audio Tool
Click here for abridged RAT User Guide.
On−line documentation for RAT is available from RAT Audio Tool.
The CD−ROM has a pdf copy of the complete RAT userguide as userguide.pdf.
2.1 What is RAT?
Unicast
A unicast
connection is like a
telephone call,
which has one
connection between
two telephones for
the duration of the
call.
Multicast
A multicast
connection can, in
principle, have an
unlimited number of
participants. If there
will be more than
two participants in
the conference, you
must use multicast.
Contact your
network
administrator to set
this up on your
LAN.
RAT (Robust−Audio Tool) is an audio conferencing tool developed
by the Computer Science Department of the University College,
London. RAT can be used for either point−to−point (unicast)
conferencing, involving a direct link between two computers, or for
multiparty (multicast) conferencing when many participants in
different locations are linked via a multicast−capable network.
2.2 Before you start
In order to receive audio, there must be either a speaker or a pair of
headphones connected to your computer. In order to transmit audio,
you must have a microphone plugged into your computer. You
should also have a ‘full duplex’ sound card installed as RAT only
supports full−duplex sound cards. It is recommended that
headphones are used in preference to speakers. Using headphones
allows you to listen and talk at the same time without interference.
Follow the installation steps as outlined in chapter 1 if RAT is not
installed already on your computer.
2.3 Starting RAT from the command line
The commands vary with platform and conference type. The
command for starting up RAT from the command line is as follows:
rat [options] hostname−or−IP−Address[/port]
In Unix this can be typed at the prompt. In Windows 95 or 98 or
NT4.0, open the RUN window from the start button and type it.
Note that on some systems, you may need to type in the full path
(see figure 2.1).
For example, to start a RAT session with someone on a computer
named asoju.oauife.edu.ng without any special options, type
rat asoju.oauife.edu.ng
The person working on asoju.oauife.edu.ng will also have to issue
the same command but with the name of your computer instead.
TTL
Figure 2.1 Starting RAT win Windows ‘95/98
TTL stands for
Time To Live and
determines how far
To start a session with a machine with IP address 140.105.36.3
using port 5006, type
the audio that you
transmit will travel.
As a guide a TTL of
15 will reach other
parts of a campus, a
TTL of 47 will
reach all parts of a
country. Another
technique called
Admin Scoping can
also be used to
control how far your
audio goes.
rat 140.105.36.3/5006
To join a multicast session on the address 224.1.1.13, setting the
TTL option to 47 and the port number to 5004, type
rat –t 47 224.1.1.13/5004
All participants in a conference must use the same port. The port
number should be an even number not less than 1024. RAT uses
both the port specified and also the next port. If no port is specified
the default port of 5004 is used. Either the hostname or the IP
address of the remote computer can be used for unicast sessions.
For multicast sessions however, it must be the IP address (usually
in the range 224.0.0.0 − 239.255.255.255). Ask your administrator
for the address to use for a particular multicast session.
2.4 Using RAT
RAT consists of a Main Window (see Figure 2.2), a Preference
window (see Figure 2.5**), and a number of user−info windows
(see section 1.3.3 below).
The RAT main window
RAT’s Main Window is split into three sections. The top section of
the main window is divided into two sections, input to the left and
output to the right. Clicking the boxes labeled Listen and Talk will
enable the devices shown below them. If you do not check the
appropriate box, you’ll be unable to hear and / or to be heard.
Beneath the input and the output device is a slider to set sound
levels which also give a numeric indication on a scale from 0 to 100
next to the device indicators. This numeric indication is useful
when setting appropriate levels with other users at the start of a
conference, providing a precise indication of your level setting.
Immediately above the slider is a graphical indication of the output
volume and the input gain.
In the middle section of the main window is a list of conference
participants with your own name at the top. Names are highlighted
whenever anyone speaks. Individual participants can be selectively
muted by clicking on their names with the middle mouse button, or
by holding down the Control Key (on the keyboard) and left
clicking on their names. Clicking on the name of a participant,
including your own, will display a Preferences panel where you can
display Personal Details as well as other technical details such as
Play Out, Decoder, Audio, and 3D Positioning by selecting the
desired category
NOTE
By default, the
output device is
Speaker and the
input device is
Microphone.
Depending on your
hardware
configuration, other
devices may be
available for input
and output. To
change between
these options, click
on the small
triangular button on
either side of the
device description.
Fig. 2.2. The Main Window of RAT
In the bottom section of the window is the session name, the
conference address and port number, the TTL and six control
buttons. From left to right, these buttons have the following
functionality:
• File control when playing from or recording to files.
• Enabling display of the reception quality matrix.
• Enabling the balloon help facility.
• Options
displays the various preferences which can be
• About
shows credits, copyright information and feedback methods.
• Quit
Closes RAT and ends the session.
2.5 Listening to others
In the main RAT Window, the slider bar at the top left is used to
adjust the volume of the speaker or headphones. When someone is
talking, the power meter above the left slider will be activated and
the name of the speaker will be highlighted in the main window.
You may need to mute your microphone while listening to others. If
you are using headphones rather than a microphone and speakers,
you can leave the microphone un−muted all the time other wise, the
microphone should only be un−muted when you wish to talk.
Possible problems with hearing others
(i) If you cannot hear other people at all:
• Make sure that the listen button is checked. If the Listen
button is not checked, the incoming audio will be muted
and the power meter will be plain black.
• Make sure that someone is actually talking. You can see
that someone is talking when the left power meter in the
RAT audio window is active and a name in the
participants’ list is highlighted.
• Check that your headset/speaker is plugged into the
computer correctly.
• Check the output volume setting. Make sure that the
volume slider for the headphones/speakers is not too far to
the left.
• Check the output type setting. Make sure that the
description of the output is correct: Headset or Speaker.
You can change between the output types available by
clicking on the small triangles on either side of the
description.
• Check that you are not running any other audio program. If
you are, you should exit from it and restart RAT.
• Check in the ‘audio’ options that your audio device is
selected.
• Check Sample rate of sender. If other participants are using
a different ‘sample rate’, then check in the Receptions
category under Preferences that a sample rate converter has
been selected.
(ii) If you can hear other people but with difficulty:
• If the other participants’ audio is too faint, ask them to turn
up their volume.
• If the audio from another participant is very broken up, the
most likely cause is either network congestion causing the
loss of their data packets, or the other user speaking too
quietly for the audio tool to distinguish the voice from the
background noise. You can tell if you are losing packets
from another participant by looking at the diamond to the
left of the participant’s name. If the diamond is green (loss
less than 5%) or orange (loss less than 10%), then the
packet loss is small so ask them to turn their volume up. If
that does not help, ask them to turn off silence suppression.
• If everyone’s audio keeps cutting out, a possible cause is
your moving windows around on the screen. When you do
this, the processing power of the computer is spent on
window management rather than audio decoding, and you
lose audio quality.
• Consider turning Lecture Mode on in the Reception
category of the Preferences menu, if you are only listening.
2.6 Speaking to others
In order to talk to other people, you must un−mute your
microphone. This can be achieved either by checking the box
labeled Talk at the top right section of the main window or or
holding down the right mouse button whilst the pointer is in the
RAT window.
If other people can’t hear you:
• Make sure that your microphone is not muted when you talk. The
microphone is muted when the Talk button in the RAT audio
window is checked and the power meter is plain black.
• If you have started from the command line, make sure that
the TTL is high enough. If your TTL is too low, you may
still be able to hear a remote participant even though they
cannot hear you. To change the TTL you will have to
restart RAT.
• You may have to disable silence suppression in the Audio
category under the Preferences menu, and use
push−to−talk.
4. Configuring RAT.
Pressing the options button in the lower part of the main RAT
window brings up a Preferences window, allowing the operation of
RAT to be modified. There are several categories of options that
can be altered and you can select anyone by using the category
menu. Press down the button in front of category and select any of
the seven options (see figure 2.3). Options can be set for the current
session by clicking on the apply button or ignored by clicking on
the cancel button.
Fig. 2.3 The Preferences Window
These are some of options you need to change often:
• If other participants complain that your audio keeps cutting
out and it is not due to packet loss, try turning up your
microphone volume. If that does not help, turn off suppress
silence ( see Box 6: Silence Suppresion). If you turn silence
suppresion off, you should mute your microphone
whenever you are speaking, otherwise you will be
transmitting background noise continuously( see Box 7:
Push to talk ) .
• If you are only listening, for example to a lecture, you
should turn Lecture Mode on. When lecture, mode is
switched on, RAT delays the playout of the audio for a
short while, allowing more time for packets to be received.
Lecture Mode is automatically off when you talk.
• You may want to change your name as it appears in the
main RAT Window. You can do that by clicking in the
Name field, editng the information . Likewise, you can
change your e−mail address, phone number and location.
These details will be available to other participants if they
click on your name in the Main RAT window.
NOTE
Some audio cards
will automatically
send audio to
headset if they are
plugged in and
speakers if not. In
this case the
description will not
toggle, but will
always display
Speakers
.
Sample Rate
This can be adjusted
in the Audio
category under
Preferences. Higher
sampling rates
improve the quality
but require greater
network resources.
Silence Suppression
.
This is used to
ensure that only
audio above a
certain level is
transmitted i.e. to
ensure that
background noise is
not transmitted.
However,
sometimes the
silence suppression
will cut out speech
as well as
background noise.
This can happen if
your microphone is
a long way from
your mouth, making
your voice sound
faint, or the
microphone is not
powerful enough.
Push−to−talk
means that you keep
your microphone
muted, and only
un−mute it when
talking. The easiest
way of achieving
this is by right
clicking anywhere
within a RAT
window, which will
toggle the Talk
button when you
talk.
MBONE Session Directory Tool (SDR)
On−line documentation for SDR is available from Session Directory Tool.
The CD−ROM has a pdf copy of the complete SDR userguide as userguide.pdf.
Linux
On Linux it may be necessary to add the route to the MULTICAST Network using the command route add
−net 224.0.0.0 netmask 240.0.0.0 gateway {place_your_ip_address_here}. The
UNIX "route" command is only available to the Super User (or system administrator).
ScientificTalk − Mathematics Discussions via Web
A profession−specific prototype tool for scientists, students and teachers to exchange information via a Web
browser, using a display of math equations in a synchronous manner. It focus on users' interests in such things as
mathematics and scientific notation. ScientificTalk is a Perl script for a standard multiway graphical Web chat.
This CGI−based application is portable across platforms and allows the viewing of occupants, sending input to
specific users, etc. While chatting on−line, it converts textual input or standard LaTeX into HTML. LLaTeX ist a
popular computer language for composing formatted scientific text for high−quality printing. It has an external
symbols keyboard as well as a composer and messenger window for the user input. There is no need for extra
plug−ins or high−speed networks. All input is passed via text−mode only.
Click here for Local User Guide and Documentation.
More information about Scientific Talk is available from http://www.collaborium.org/ScientificTalk.
[Top] [Contents] [Index] [ ? ]
ScientificTalk − Table of Contents
1. Introduction
1.1 What is ScientificTalk?
1.2 Why a profession−specific tool for e−collaboration?
1.3 Typical types of e−collaboration
1.4 Limitatitons of ScientificTalk
2. Installing ScientificTalk
2.1 Where to get ScientificTalk?
2.2 Installing ScientificTalk under Linux
2.3 Installing NetPlot
2.4 ScientificTalk Administration
3. Usage
3.1 How to Configure your Web Browser
3.2 Entering a ScientificTalk Session
3.3 Math Discussions writing LaTeX Input
3.4 Example of Complex Math Formulae
3.5 Seeing All LaTeX Inputs
3.6 Not Familiar with LaTeX − External Equation Editor
3.7 The Equation Viewer
3.8 NetPlot Input/Output
4. Troubleshouting Help and Info
4.1 Articles
4.2 Virtual Community Standards
This document was generated by Canessa Enrique on May, 18 2001 using texi2html
ScientificTalk GUIDE
1. Introduction
3. Usage
[<] [>]
[ << ] [ Up ] [ >> ]
1. Introduction
In an increasingly interconnected world, electronic communication tools play an essential role in facilitating the
exchange of ideas and discussions over the Internet. In particular, highly needed are profession−specific tools for
synchronous distance collaboration that focus on users' interests, such as mathematics and science.
As early as 1990, when the original view of the Web emerged, see e.g.,
http://www.w3.org/History/1989/proposal.html
scientists at CERN realized it would be extremely valuable to have a readable and writable collaborative medium.
Today's needs of the scientific community are far from met with regard to the use of mathematical notation and
symbols on−line, which is part of the fabric of scientific professional life. Mathematical notation on the Web is a
completely open technological challenge specially with the application of the new standard for MathML (i.e.,
Mathematics Markup Language proposed by the W3 Consortium −http://www.w3.org/Math).
ScientificTalk aims to contribute to this challenge.
[<] [>]
[ << ] [ Up ] [ >> ]
ScientificTalk is a profession−specific prototype tool for scientists, students and teachers to exchange
information via a web browser, using a display of math equations in a synchronous manner. It focus on users'
interests in such things as mathematics and scientific notation. and follows an early goal for the Web to be a
readable and writable collaborative medium.
ScientificTalk allows users to interactively:
• Enter letters and numbers (data) from the keyboard (similarly to a Web Chat room or to the Talk Unix
command) by writing shared text.
• To select Mathematics signs/symbols from a Equation editor to write down, display and share Equations,
Formulae, etc.
• To enter La@TeX dialect input (if your are familiar with it). La@TeX is a popular computer language for
composing formatted scientific text for high−quality printing.
For those not familiar with La@TeX commands, ScientificTalk offers an external symbols keyboard as
well as a composer and messenger window for the user input. There is no need for extra plug−ins or high−speed
networks −all input is passed via text−mode only. On the client side, Netscape (v4.0 or greater) needs a simple
character−set configuration see section 3.1 How to Configure your Web Browser. ScientificTalk under MS
Internet explorer runs without any specific Web Browser configuration for "symbols" fonts.
ScientificTalk is written in Perl with the added feature to convert standard input (by selecting symbols of
typing La@TeX commands directly on your browser) into HTML for the display of Mathematics (immediately
after automatic page refreshing).
While chatting on−line, it converts textual input or standard La@TeX into HTML. The math displayed on the
browser is rich because of the La@TeX typesetting and Ian Hutchinson's powerful TeX−to−HTML translator, tth,
available at http://hutchinson.belmont.ma.us/tth Very important, tth produces Web documents that are compact
and managable and faster−viewing! It translates the equations instead of converting them into images.
A simple browser set−up is only needed on your computer. There is NO NEED for extra plugs−in or fast
bandwidth connectivity.
[<] [>]
[ << ] [ Up ] [ >> ]
[<] [>]
[ << ] [ Up ] [ >> ]
The hope is to offer scientists a tool to communicate their own scientific results and achievements to their peers
world−wide.
The aim of ScientificTalk prototype package is to fill up a gap for a professional Web Chat Application
with scientific language capabilities to communicate (in two/multi syncronous ways) with remote Physicists,
Mathematicians, Engineers, Students, etc.
Indeed, the large tag repertoire of the HTML 4.x language does not cater to mathematics since they cannot mark
up complex mathematical expressions. Usually, to create technical documents with mathematical or scientific
content, web authors resort to methods involving images (e.g., screen captures of equations), which means the
sharing of scholastic and scientific material by lecturers, students, etc. is often a many−step process. There are a
few available applets and plug−ins that can render MathML in a browser (which are not necessarily designed for
synchronous collaboration). The Mathematical Markup Language, MathML, is a recommendation of the W3C,
which provides a foundation for including mathematical expressions in web pages. (Some applets and plug−ins
are currently available, which can render MathML in a browser.) See W3C at http://www.w3.org/Math for a
complete list of technical/scientific document viewers and renderers such as the Scientific MessageBoard
WebEQ, the IBM techexplorer, EzMath editor and La@TeX2HTML.
[<] [>]
[ << ] [ Up ] [ >> ]
More opportunities for learning and growth are available if we can share ideas via a computer environment that is
responsive to our professional needs. For example, using simplified scientific notation on the Web can lead to
faster, more effective results. Electronic tools, designed for collaboration and based on open source technologies,
will continue to play an important role in an globalized world.
Typical types of e−collaboration, besides the most popular simple person−ro−person e−mails exchanges, are:
Distribution lists and/or MessageBoards (e−forums, e−seminars and e−Conferences) based on e−mail archiving in
a Web server coming from many sources, and the popular text−based web chats using Java applets and CGIs. As
an application of the Extensible Markup Language (XML) and with adequate style−sheet support, MathML will
ultimately make it possible for browsers to natively render math expressions, including threaded on−line
discussions.
[<] [>]
[ << ] [ Up ] [ >> ]
Although the ScientificTalk prototype has proven that it is possible to carry out synchronous math
discussions on the Web between distant clients today, our to−do list is still long. For example, it would be useful
to save a complete session as a La@TeX file (in order to restart an on−line discussion from a given session or
collaboratively write La@TeX documents on the Web), display plots from a given function, create small
transparent .gif files, and extend its language capabilities to symbols in other domains, such as chemistry.
Presently, the ScientificTalk application runs using CGI which is portable across platforms and languages.
CGI is more adequate for lightly loaded machines since CGI is inherently slow, requiring the Web server to create
a new process each time an external program is invoked. On increasing usage, switching to modular Perl −in a
possible future release of ScientificTalk− to attach a program (subroutine) to the server process would
allow maximum performance.
[<] [>]
[ << ] [ Up ] [ >> ]
ScientificTalk can be freely installed and used for non−commercial purposes only. Version 2.2 and highers
also incorporates NetPlot −a public domain package for plotting using 'gnuplot' (from
ftp://ftp.dartmouth.edu/pub/gnuplot) and the ImageMagick tool 'convert' (available at
http://www.wizards.dupont.com/cristy/ImageMagick.html) for the output display of
NetPlot.
[<] [>]
[ << ] [ Up ] [ >> ]
The Perl code sources of the ScientificTalk program are Freely Available provided that you duplicate all of
the original copyright notices and associated disclaimers of all other packages used. These can be downloaded
from the Web site
http://www.collaborium.org/ScientificTalk
[<] [>]
[ << ] [ Up ] [ >> ]
To install ScientificTalk, you need
• Perl interpreter version 5 or higher.
• compile 'tth' in the directory 'tth_C/'
gcc −o tth tth.c
• In the file 'sci−talk.cgi', change to your Path (i.e., directory containing the ScientificTalk
file) and any other relevant lines afterwards;
$setup_path = "/var/www/cgi−bin/sci−talk";
$web_path = "/var/www/htdocs";
(or whatever is your path!)
Check also your path for Perl in 'sci−talk.cgi':
#!/usr/bin/perl
• You must set permitions on Sci_Talk and Sessions directories and allow ALL to write in it.
drwxrwxrwx
drwxrwxrwx
Sci_Talk
Sessions
• Copy the 'Images/*.gif' files into your Web server directory
/var/www/htdocs/images
[<] [>]
[ << ] [ Up ] [ >> ]
NetPlot is a Web interface for the public domain program 'gnuplot' using the ImageMagick tool 'convert'.
It was designed primarily for students of engineering and natural sciences by D. Krstinic & I.Slapnica (1999) as
indicated in the on−line display.
NetPlot produces images in gif format which may be found in Netplot1.1/tmp directory and it should
work on any Linux box. You must set permitions on tmp directory and allow ALL to write in it. Then proceed as
follows
• Edit 'c_source/Makefile' to your settings
# For gcc
CC= gcc
CFLAGS= −g −Wall
RM= /bin/rm −f
CGIDIR= /home/httpd/cgi−bin/
# Ovo mi treba za test
BINDIR= bin/
(Note the "/" at the end)
• Type 'make'
• Two executables CGI will be automatically installed in CGIDIR:
NetPlot.cgi
formiraj.cgi
• Type 'make clean'
• After compilation, move the whole directory 'Netplot1.1' to your Web server path, e.g.,
/var/www/htdocs/NetPlot1.1
Then, you can delete the directory
/home/httpd/htdocs/NetPlot1.1/c_source
• Check also the files:
NetPlot.c:
#define SLIKA_FILE "/home/httpd/htdocs/NetPlot1.1/tmp/Fig\0"
and
formiraj.c:
#define CLEAR_DIR "rm −f /usr/local/httpd/htdocs/Netplot1.1/tmp/*\0"
... That should make it!
[<] [>]
[ << ] [ Up ] [ >> ]
WARNINGS:
• Since ScientificTalk waits automatically for new input using Push Web technologies (to then
refresh), it is recommended to use a Web server Timeout (i.e., the number of seconds before receives and
sends time out) of about:
Timeout 18000
(on the Web Apache server modify the 'httpd.conf' file).
• Avoid using ScientificTalk with Proxies allowed (since Proxies also have a TimeOut configuration
parameter).
• From time to time clean up the directories
Sessions/
Sci_Talk/
This is a necessary step! If people do not "LOG OFF" using the ad−hoc button, such sessions remain
stored.
Also from time to time delete old 'see_LaTeX_xxxxxx.html' files. In addition, with
ScientificTalk−version 2.2 and higher you also need to clean
NetPlot1.1/tmp
[<] [>]
[ << ] [ Up ] [ >> ]
3. Usage
ScientificTalk is easy and intuitive to use!
[<] [>]
[ << ] [ Up ] [ >> ]
On the client side, Netscape (v4.0 or greater) needs a simple character−set configuration. ScientificTalk
under MS Internet explorer runs without any specific Web Browser configuration for "symbols" fonts.
[<] [>]
[ << ] [ Up ] [ >> ]
Just enter the username field to be associated to you.
[<] [>]
[ << ] [ Up ] [ >> ]
You can enter letters and numbers (data) from the keyboard (similarly to a Web Chat room or to the Talk Unix
command) by writing shared text including La@TeX dialect input (if your are familiar with it) as illustrated next.
Just press the SUBMIT Button to send your input (in−line or centered) to your Colleagues on the net. By default
your input (in ASCII mode) is sent to ALL occupants of the chat session. You can send your message to a
particular Occupant by filling his/her username in the ad−hoc form field. You may want also to preview your
input before send it by selecting the 'Preview first' button. The Occupants list is available by selecting the
'View Occupants' Button. At any time, you may refresh the current Web session as well.
[<] [>]
[ << ] [ Up ] [ >> ]
ScientificTalk can do this and even more complex Math Formulae in real time!
[<] [>]
[ << ] [ Up ] [ >> ]
The La@TeX dialect input from your Colleagues at the session can be obtained by pressing the 'See LaTeX
Input' Button close to the Message Numbering.
[<] [>]
[ << ] [ Up ] [ >> ]
For those not familiar with La@TeX commands, ScientificTalk offers a symbols keyboard as well as a
composer and messenger window for the user input. You can select Mathematics signs/symbols from the external
Equation Editor:
[<] [>]
[ << ] [ Up ] [ >> ]
After you select Math signs/symbols from the external Equation Editor, ScientificTalk writes down and
display your input as illustrated below. By pressing the Load Button you can share Equations, Formulae, etc with
others connected to your session.
[<] [>]
[ << ] [ Up ] [ >> ]
NetPlot is also available via ScientificTalk as an extra facility to plot the Formulae being discussed.
Give an input as shown below.
[<] [>]
[ << ] [ Up ] [ >> ]
The code sources of the ScientificTalk program are written/adapted and maintained by Enrique Canessa
and are Freely Available provided that you duplicate all of the original copyright notices and associated
disclaimers of all other free packages used. This software can be freely used for non−commercial purposes. The
Author of ScientificTalk, will appreciate hearing from you, to know your experiences, comments,
suggestions, etc on this prototype tool to facilitate profession specific distance collaboration. If you want to get
updated info on ScientificTalk, Troubleshouting Help, FAQ, etc please visit the Web site
http://www.collaborium.org/ScientificTalk
or send an e−mail to
canessae@collaborium.org
4.1 Articles
[<] [>]
[ << ] [ Up ] [ >> ]
4.1 Articles
• 'Low−Bandwidth Communication Tools for Science'
by Enrique Canessa and Clement Onime
Linux Journal, #75 July 2000.
• 'Enabling Synchronous Math Discussions on the Web'
by E. Canessa
Computing in Science & Engineering, Vol. 1, No. 5, Sep/Oct 1999.
[<] [>]
[ << ] [ Up ] [ >> ]
ScientificTalk offers a virtual place for interactive communication and free exchange of ideas with other
members of the scientific community. You, the user, are responsable for content and style. So, be forewarned!
Like the community you live in, this virtual community works best when everyone follows basic standards of
conduct. Use common sense. Your privilege to use and contribute depends on your compliance with the following
simple guidelines:
• Don't write personal attacks on other members or other offensive, defamatory or otherwise inappropriate
material.
• The use of ScientificTalk is for lawful purposes only and at user's sole risk. Your conduct is your own
responsability.
All requests, usage or abuse of any ScientificTalk server are stored in log files.
... Have a nice ScientificTalk with your Colleagues!
About this document
The buttons in the navigation panels have the following meaning:
Button
Name
Go to
[<]
Back
previous section in reading order
[>]
Forward next section in reading order
[ << ]
FastBack previous or up−and−previous section
[ Up ]
Up
up section
[ >> ]
FastForward next or up−and−next section
[Top]
Top
cover (top) of document
[Contents] Contents table of contents
[Index]
Index
concept index
[?]
About
this page
From 1.2.3 go to
1.2.2
1.2.4
1.1
1.2
1.3
where the Example assumes that the current position is at Subsubsection One−Two−Three of a document of the
following structure:
• 1. Section One
♦ 1.1 Subsection One−One
◊ ...
♦ 1.2 Subsection One−Two
◊ 1.2.1 Subsubsection One−Two−One
◊ 1.2.2 Subsubsection One−Two−Two
◊ 1.2.3 Subsubsection One−Two−Three
◊ 1.2.4 Subsubsection One−Two−Four
♦ 1.3 Subsection One−Three
<== Current Position
◊ ...
♦ 1.4 Subsection One−Four
Windows
The wintalk application provides it's own help pages. Which may be accessed when the application is running.
Linux
The UNIX talk client is executed after opening a new xterm (terminal) window. For more information try the
following command "man talk" from the UNIX prompt.
VIC video Tool
Click here for abridged VIC User Guide.
On−line documentation for VIC is available from VIC Video Tool.
The CD−ROM has a pdf copy of the complete VIC userguide as userguide.pdf along with an FAQ as faq.pdf.
3.1 What is VIC?
Unicast
VIC is a video conferencing application developed by the Network
Research Group at the Lawrence Berkeley National Laboratory in
A unicast
collaboration with the University of California, Berkeley. The included
connection is like a version of VIC is that modified by the Computer Science Department
telephone call,
of the University College, London. VIC can be used for either
which has one
point−to−point (unicast) conferencing, involving a direct link between
connection
two computers, or for multiparty (multicast) conferencing when many
between two
participants in different locations are linked via a multicast−capable
telephones for the network.
duration of the call.
3.2 Before you start
Multicast
No special hardware is need to receive and display video.
A multicast
connection can, in In order to transmit video, you must have a camera and a frame
principle, have an grabber connected to your computer.
unlimited number
of participants. If Follow the installation steps as outlined in chapter 1 if VIC is not
there will be more installed already on your computer.
than two
participants in the 3.3 Starting VIC from the command line
conference, you
must use multicast. The commands vary with platform and conference type. The command
Contact your
for starting up VIC from the command line is as follows:
network
administrator to set vic [options] hostname−or−IP−Address/port
this up on your
LAN.
In Unix this can be typed at the prompt. In Windows 95 or 98 or
NT4.0, open the RUN window from the start button and type it. Note
that on some systems, you may need to type in the full path (see figure
3.1).
Frame Grabber
For example, to start a VIC session with someone on a computer
A frame grabber named asoju.oauife.edu.ng and port 7002 without any special options,
is a piece of
type
hardware which is
installed in the
vic asoju.oauife.edu.ng/7002
machine and
translates the data The person working on asoju.oauife.edu.ng will also have to issue the
transmitted by the same command but with the name of your computer instead.
camera in to the
correct encoded
format for VIC, or
some other video
application.
Figure 3.1 Starting VIC in Windows ‘95/98
To start a session with a machine with IP address 140.105.36.3 using
port 7002, type
vic 140.105.36.3/7002
To join a multicast session on the address 224.1.1.13, setting the TTL
option to 47 and the port number to 7002, type
vic –t 47 224.1.1.13/7002
All participants in a conference must use the same port. The port
number should be an even number not less than 1024. VIC uses both
TTL stands for
the port specified and also the next port. Either the hostname or the IP
Time To Live and address of the remote computer can be used for unicast sessions. For
determines how far multicast sessions however, it must be the IP address (usually in the
the audio that you range 224.0.0.0 − 239.255.255.255). Ask your administrator for the
transmit will travel. address to use for a particular multicast session.
As a guide a TTL
of 15 will reach
3.4 Using VIC
other parts of a
campus, a TTL of VIC consists of a Main Window (see Figure 3.2), a Preference window
47 will reach all (see Figure 3.5**), and a number of user−info windows (see section
parts of a country. 1.3.3 below).
Another technique
called Admin
The VIC main window
Scoping can also
be used to control When nobody is transmitting video on the address/port that VIC has
how far your audio been started up on, the main window displays a “Waiting for video…”
goes.
message (see Figure 4). Otherwise, the Main VIC Window will
automatically start displaying small thumbnail versions of the video
images that are currently being transmitted. The default size of the
Main VIC Window can display three thumbnail images, so in order to
see more thumbnail images you must enlarge the Main VIC Window.
TTL
Fig. 3.2. The Main Window of VIC
Each thumbnail image can be enlarged by clicking on it once. A larger
image will then be displayed in a separate window. This larger image
can be resized using the Size… button at the bottom the window.
• Identification Text. The identification text contains the name of the
NOTE
By default, video is
not transmitted
automatically.
participant, information about the machine that the participant is using
(either the full Internet name or the IP address) plus information about the
video format that the participant is transmitting. The name appears
greyed−out if a session message has not been received from that site for a
number of seconds (each VIC tool sends out session messages
periodically). A greyed−out name usually indicates that the site has lost
connectivity or that VIC has been abnormally aborted or stopped.
• Below the identification text, there is information about the video stream
that you are receiving from that participant. It tells you how many frames
per second and how many kilobits per second you are receiving. These
numbers may differ from the actual sender’s rate because of network loss,
and may differ from the local display rate because of CPU overload. The
number in brackets tells you the packet loss from that participant on the
video stream.
• mute button. When the mute button is selected, VIC will no longer decode
video from the corresponding participant. The image will appear frozen on
your screen. You can mute your own image as well, but remember that
even though it appears frozen on your own screen, it will not appear frozen
on other participants’ screens unless they mute your image at their end. It is
advisable that, in order to limit the load on your machine, the images that
you are not interested in are muted.
• color button. The color button controls the colour options of the received
video. By default, video is displayed in colour but it can also be displayed
in greyscale. Using greyscale reduces the CPU load on machines without
TrueColor displays because colour dithering is costly in terms of CPU
loading. The color button does not affect your transmitted video.
• Info button. Clicking on the info button will give you a menu containing
the following items: Site info, RTP Stats, Decoder Stats, Mtrace from,
Mtrace to. Site info will open a window containing information about the
participant, such as name, e−mail address, video format, version of VIC
etc. (see Figure 5). From this window you can press the Stats button which
will bring up a small menu containing two items, RTP and Decoder.
Selecting RTP summons a window containing the RTP statistics for that
participant. Selecting Decoder displays the decoder statistics for that
participant. The RTP statistics window consists of a grid: The grid contains
three columns. The first column, labelled EWA (Exponentially Weighted
Average), is the change since the last sampling period (i.e., change over the
last second); the middle column, labelled Delta, is a smoothed version of
the EWA and the last column, labelled Total, is the accumulated value
since start−up. Clicking any of the buttons in the left column opens a Plot
Window displaying the statistics for that parameter graphically. The
windows containing RTP statistics and decoder statistics can also be
opened directly from the info menu by selecting RTP Stats and Decoder
Stats respectively.
In the bottom section of the main window are three buttons which have
the following functionality:
NOTE
1. Menu displays the various preferences which can be modified
2. Help shows a one page quick help screen.
3. Quit Closes VIC and ends the session.
Video transmission
may be bandwidth
2.6 Viewing video from others
intensive.
Acceptable quality
(low speed) video In the main VIC Window, click on the thumbnail image of the video
is possible with a you wish to enlarge.
setting of between
Possible problems with receiving video
15 to 30 kbps.
• If you cannot see all the participants in VIC: Make sure that
the VIC video window is large enough to accommodate all the
Transmission Rate
This can be
adjusted using the
Control Rate slider.
Higher
transmission rates
improve the quality
but require greater
network resources.
Video Mute.
pictures. Depending on your window manager, you can
normally enlarge the VIC video window by dragging the edges
of the window. If you still cannot see them, it may be because
they are not transmitting video.
• If you cannot see yourself in VIC: Make sure that you are
transmitting video. There should be a little red square on the
Transmit button in the menu which you get by pressing the
Menu button in the VIC video window. If not, press Transmit.
Also make sure that the camera’s power supply is plugged into
the mains and that the camera itself is connected to the
computer.
Decoding multiple 3.6 Sending video to others
video streams can
be CPU intensive. In order to transmit video to other people, you must enable
Muting some video transmission. This can be achieved by clicking the Menu button at the
bottom section of the main window.
sources can
dramatically
4. Configuring VIC.
improve
performance on
Pressing the Menu button in the lower part of the main RAT window
slow machines.
brings up a Preferences window, allowing the operation of VIC to be
modified.
Fig. 3.3 The Menu/Preferences Window
These are some of options you need to change often:
• Transmit and Release buttons. Click once on Transmit to
transmit video. To stop transmitting, either click Release or
click Transmit again. When you stop transmitting, your image
will freeze. You can get rid of frozen pictures by moving the
mouse pointer onto the thumbnail image that you would like to
close down, and pressing the letter d on the keyboard. You
should not click on any of the mouse buttons.
• You can change your name as it appears next to your
thumbnail image by editing the Name field: click in the field,
edit the text and press the Enter key on the keyboard.
• If you want to see a list of all the people running VIC on the
address/port, click the Members button.
• If you need to confirm the address, port, ID and TTL of VIC
for the session, all of this information can be found in the VIC
Menu Window.
• You can adjust the maximum amount of bandwidth (kbit/s)
that you are using and the maximum number of frames per
second (fps) that you a transmitting by adjusting the slider bars
to the right of the Transmit and Release buttons in the VIC
Menu Window.
VNC Shared Desktop Tool
Click here for the locally installed Documentation.
On−line documentation for VNC is available from Virtual network Computing.
The CD−ROM has a pdf copy of the complete VNC userguide as vnc.pdf along with an FAQ as
frequently_asked_questions.pdf.
Some recent changes:
Windows packages updated to 3.3.3r9 − 19/3/01
No virus in WinVNC! − 8/3/01
A list of all known encoding numbers in use − 8/11/00
Unix packages updated to 3.3.3r2 − 26/10/00
Note: The FAQ and some other bits of the documentation are
constantly being updated. We only record major changes here.
What is VNC? − A practical introduction
VNC stands for Virtual Network Computing. It is, in essence, a remote display system
which allows you to view a computing 'desktop' environment not only on the machine
where it is running, but from anywhere on the Internet and from a wide variety of
machine architectures.
The VNC system allows you to access the same desktop from a wide
variety of platforms.
Many of us, for example, use a VNC viewer running on a PC on our desks to display
our Unix environments which are running on a large server in the machine room
downstairs.
What makes it different from other systems?
For this simple mode of operation, you could achieve a similar effect by installing an X
server on your PC. The important factors which distinguish VNC from other remote
display systems such as X are as follows:
• No state is stored at the viewer. This means you can leave your desk, go to
another machine, whether next door or several hundred miles away, reconnect
to your desktop from there and finish the sentence you were typing. Even the
cursor will be in the same place. With a PC X server, if your PC crashes or
is restarted, all the remote applications will die. With VNC they go on
running.
• It is small and simple. The Win32 viewer, for example, is about 150K in size
and can be run directly from a floppy. There is no installation needed.
• It is truly platform−independent. A desktop running on a Linux machine may
be displayed on a PC. Or a Solaris machine. Or any number of other
architectures. The simplicity of the protocol makes it easy to port to new
platforms. We have a Java viewer, which will run in any Java−capable
browser. We have a Windows NT server, allowing you to view the desktop of a
remote NT machine on any of these platforms using exactly the same viewer.
(The NT server is not multi−user − see the documentation). And other people
have ported VNC to a wide variety of other platforms. Click the 'Contributed'
button on the left for details.
• It is sharable. One desktop can be displayed and used by several viewers at
once, allowing CSCW−style applications.
• It is free! You can download it, use it, and redistribute it under the terms of the
GNU Public License. Both binaries and source code are available from the
download page, along with a complete copy of this documentation.
Where does the name come from?
The name originates from our development of very−thin−client ATM network
computers. The Videotile was essentially an LCD display with a pen input and a fast
ATM connection. Because the VNC viewer is a software−only version of this 'ATM
Network Computer', and so provides 'workstations' which can be created or deleted at
will, we named the system Virtual Network Computing.
Can I see what VNC looks like?
We have some screenshots of very simple VNC desktops running and being displayed
on a variety of platforms.
Follow the links on the left to find out more...
For comments, feedback, etc, please see the 'Keeping in touch' page.
Copyright 1999 − AT&T Laboratories Cambridge
VNC under new ownership!
In January 1999, AT&T acquired the lab founded 12 years earlier as Olivetti Research
Laboratory, and more recently known as ORL.
VNC is therefore now a project of AT&T Laboratories Cambridge, and we look forward
to its continued development and the new opportunities that this partnership presents.
VNCviewer for Windows CE 2.x
Now you can view X Windows applications on your PDA!
The VNC viewer for Windows CE is based heavily on the Win32 viewer. It requires
Windows CE 2.0 or later.
To install, make sure that TCP/IP is working correctly on your WinCE machine (try
browsing the web), and we suggest you set the communications link to the fastest speed
it can manage. Try connecting to the smallest desktop you can, initially. If you're
viewing a PC or Mac, set the server's screen to 640x480 or so and 256 colours. If you're
connecting to an Xvnc server, then you can of course create a desktop of any size you
like. Larger screens will work − I've viewed a 1280x1024x24 screen here, but the initial
screen download takes quite a while, and in this version a copy of the entire remote
screen is kept in the PDA's memory, so you'll need plenty of RAM free!
Copy the appropriate vncview.exe binary onto the PDA and run it. If you have used the
Windows viewer the dialogs and menus should look very familiar. You might want to
use the Options button in the Connection dialog to request 8−bit pixels only. You can
set this from the command line, so I suggest you create a shortcut to the vncview
executable and include the '−8bit' option in the target command line.
Most of the command−line arguments and options are taken directly from the Win32
viewer, and you should look at the Windows viewer documentation for more details.
The Windows CE viewer does not operate in full−screen or listening mode. There is
no console−mode debugging, and some of the other Win32 viewer options are not
implemented.
This is certainly the smallest X terminal I've ever used!
go back to documentation
VNC Contributions
Many people have contributed to the VNC project, with ports to new platforms, and
various other modifications and extras.
We have divided the contributions into two pages:
New platforms
Modifications
and add−ons
VNC documentation
Note: A detailed paper about VNC can be found as:
Tristan Richardson, Quentin Stafford−Fraser, Kenneth R. Wood & Andy
Hopper, "Virtual Network Computing", IEEE Internet Computing, Vol.2
No.1, Jan/Feb 1998 pp33−38.
You can download it in Acrobat format here (760k).
The following documentation assumes a basic familiarity with the terms and components
used in VNC. See 'What is VNC?' and 'Getting Started' for introductory information,
and the Frequently Asked Questions for common queries.
For a quick overview, you can download versions of the VNC Video.
See also 'What's new in the VNC packages?'.
Technical documentation − Table of Contents
How VNC works
VNC servers
♦ VNC server for X (Unix)
♦ WinVNC (VNC server for Windows)
♦ MacVNC (VNC server for PPC Macintosh)
♦ rfbcounter (A really dumb server!)
VNC viewers
♦ VNCviewer
♦ VNCviewer
♦ VNCviewer
♦ VNCviewer
♦ VNCviewer
for X
for Windows
for Java
for Macintosh
for Windows CE
The VNC protocol
Making VNC more secure with SSH
Internal AT&T extensions
Note: This complete VNC documentation is available as a downloadable package.
Downloading VNC
The VNC system is available for general use under the conditions of the GNU General
Public Licence. You should be aware of the terms and conditions of this licence, which
is also contained in the distribution itself.
Please fill in the form below to download VNC. It's not required, but we'd be grateful if
you filled it in. We're just curious about who's using it and what they're using it for. This
information is purely for our own VNC−related use and will not be disclosed to any
other party. Neither will we, in general, send any mail to any address given here. If you
want to keep in touch with VNC developments after downloading, please join the
mailing list.
This page is part of the downloadable documentation; if you are looking at a local copy,
please check the latest online version − things may have changed!
There is some other great free software which is downloadable from our site.
VNC download
Your Name:
Your email address:
Your organisation:
Comments:
We would be interested to
hear any comments about
your interest in VNC, your
expected use, how you heard
about it, etc.
Binary packages:
These include the server and viewer for a particular platform. They all include the
Java viewer. Each binary package is (very approximately) 1Mbyte in size, and the
tgz version is usually the smallest. See the release history page for details of
recent changes.
If your platform isn't listed here, somebody else may have ported VNC to it. Have a
look at the contribs page.
Please choose the package(s) of interest. You can always come back for more later!
Linux 2.x for x86
(3.3.3 is glibc only)
Solaris 2.5 (SPARC)
Windows 9x/2000/NT (Intel Win32)
Macintosh
Beta
− 68K processor
NB: Stuffit 5.5 needed to unpack!
(Beta viewer)
− PPC processor
NB: Stuffit 5.5 needed to unpack!
(Beta)
DEC Alpha OSF1 3.2
Windows CE 2.x
Beta viewer (50K)
− SH3 processor
− MIPS processor
Other packages:
All sizes given are very approximate and meant as a guide only.
Copy of the online documentation (500K)
Unix sources (2.5M)
see also contribs
Windows sources (250K).
Macintosh sources (800K).
Java sources (30K)
Windows CE sources (100K)
The source code of a really simple VNC server is also available from the rfbcounter
page.
VNC Extras, Add−ons and modifications
Firewalls
Restricting connections by IP address (TCP wrapper)
Zlib compression
x2vnc
Enhanced Java viewer for JDK 1.1
VncMonitor
VNCProxy
VNConnector
VNCScan
iXvnc
Share a Single Window
Wallpaper disabling version of WinVNC
Proxy for session capture/replay
VNC Tight Encoder
x0rfbserver − accessing a standard X server with VNC
Dual Monitors in Windows
Policy Editor Template for Windows
Fastpush − Batch File Installer for Windows NT
IPv6 patch for X VNC
Firewalls
Karl Harkimian <hakimian@aha.com> sent in a patch to compile the Unix viewer so
that it can reach external servers through a SOCKS firewall. It's very simple, and the
patch is here: socks−patch.txt. These patches apply to 3.3.1 but they are simple enough
to apply by hand to later versions.
Obviously, you cannot access a server inside a firewall from outside without specially
configuring the firewall; that's what firewalls are for! Christian A. Lademann cal@zls.de
has used an elaborate scheme:
I have had a problem with connections from my workstation to a PC connected to a
customers LAN, because my firewall as well as the customers firewall does
IP−masquerading. This means that my workstation can only connect to the customers
firewall and the customers PC can only connect to my
firewall. The reason for this is, of course, to hide the complete LAN behind the
firewall(s).
Christian's solution involved a script listening on a single accessible well−known port
behind the firewall, which could accept incoming connections and redirect them to the
appropriate machine. His full details are here: rvnc.txt
Harmen van der Wal has made a patch to the VNC 3.3.3 Java Viewer Source
Distribution, that allows the viewer to connect through a firewall using an HTTP proxy.
This patch is available at http://www.xs4all.nl/~harmwal/vnc/readme.html
Restricting connections by IP address (TCP wrapper)
Note that similar functionality is now built in to the Windows VNC server.
Wolfram Gloger <wmglo@dent.med.uni−muenchen.de>.
writes:
In the list archives I notice there was a discussion of implementing access control for
Xvnc with the tcp wrapper library. I have now implemented this for 3.3.2r2. You will
need to have tcpd.h and libwrap already installed.
Wolfram's full message is at
http://www.uk.research.att.com/vnc/archives/1998−09/0168.html .
Jared Smolens <jsmolens@andrew.cmu.edu> has done an equivalent for WinVNC. He
writes:
... I wrote some code to read a list of IPs to allow and IPs to deny from a text file. The
rules are identical (to the best of my knowledge) to /etc/hosts.allow and /etc/hosts.deny
on my Linux box and the text file's syntax is close to that format. This is a feature which
has been discussed at least once on the mailing list, so I thought that you might want to
add it to the official code base.
The x86 binaries and source code are available at:
ftp://wik.res.cmu.edu/pub/vncip_bin.zip and ftp://wik.res.cmu.edu/pub/vncip_src.zip
I have two new files, ipauth.h and ipauth.cpp. I made some changes to vncclient.cpp (but
not the header) to use the class and disconnect unauthorized clients. I also have a sample
"iplist.txt" file which contains the allowed IPs.
The format of the text file works like this:
<ALLOW|DENY> <Partial/full IP>
<DENY ALL>
An unlimited, unordered list of IPs (or partial IPs) may be entered into the file like this:
ALLOW 128.2.93.
ALLOW 128.2.87.80
DENY 128.220.
DENY ALL
In this case, the DENY 128.220. is redundant because of the DENY ALL, but you get the
point. ALLOW ALL is the default, and if the user specifies that, it is ignored. Allows
always take precidence over denies. This code is not case sensitive.
I am fairly sure that I got rid of all of my memory leaks (I ran Purify on it, but I have
done some slight modifications since then). I also use the fstream library. I don't know if
you consider this to be too
much overhead.
−− Jared Smolens
zlib compression
Dave DeBarr (debarr@mitre.org) has modified the X server and viewer to use
zlib−based compression. We plan to incorporate something similar in the standard
release before long, but until then you can find his patches at:
http://www.uk.research.att.com/vnc/archives/1998−08/0039.html
In addition, Dave has provided patches for the Windows viewer at:
http://www.uk.research.att.com:80/vnc/archives/1998−08/0228.html
Luis B. Almeida has also created a version of the Windows software which you can get
from ftp://neural.inesc.pt/pub/lba/vnc .
Warren Toomey has done an alternative implementation of zlib compression, along with
some other hacks, which you can get from:
http://minnie.cs.adfa.edu.au/VNC_bits/
x2vnc
Here's a different twist to VNC. Fredrik Hubinette hubbe@hubbe.net has written a
VNC−based variation on the popular x2x program. If you run x2vnc on an X server,
you can move off the side of the screen and the mouse movements will then be sent to a
VNC server (eg. a PC sitting beside it) He writes:
x2vnc is basically a stripped down version of the vncviewer but with slightly different
goals and a very different GUI.. :)
x2vnc emulates a 'dual head' setup by catching when the user tries to move the pointer
past the edge of the screen. This allows me to control
both computers from one mouse/keyboard.
I have made x2vnc available for download from my web site:
http://www.hubbe.net/~hubbe/x2vnc.html
Enhanced Java viewer for JDK 1.1
Muddassar Farooq <mfarooq@cse.unsw.edu.au> has produced an enhanced version of
the Java VNC viewer.
When run as a standalone application, it adds scrollbars, and the ability to specify
'host:display' as with the other VNC viewers.
It uses the Java 1.1 event model, so is 'better' Java, but won't run on older JVMs &
browsers.
You can get it from http://www.uk.research.att.com/vnc/contrib/mfviewer.tar.gz .
VncMonitor
John Wilson <tug@wilson.co.uk> writes:
VncMonitor is intended for those people who need to monitor several remote systems. A
single window is used to present all the displays. The tab or backtab key allows the user
to switch between systems. The return key causes the currently viewed system display to
be transferred to its own window and the user can interact with the system using the
mouse and keyboard. Closing
the new window returns the monitored system display back to the initial window.
The configuration of VncMonitor is controlled by a file which contains all the
information about what systems are to be monitored.
A version can be downloaded from:
http://www.wilson.co.uk/Software/vnc/VncMonitor.htm
VNCProxy
John Wilson has also written VNCProxy, a Java proxy that can sit between a VNC viewer
and server and can be configured to map various incoming requests to different VNC
servers. You can find it at http://www.wilson.co.uk/Software/vnc/proxy/VncProxy.htm
VNConnector
Alastair Burr has written a similar program to allow you to switch between multiple VNC
servers. VNConnector can be found here.
VNCScan
Steve Bostedor has written a utility which will scan a network and list the machines
running VNC servers. You can find it at http://tgcs.web−it.com. You'll need to email
Steve for the password to unzip the file.
(You can also use port scanners such as nmap or Netcat to scan for the ports used by
VNC. See the VNC FAQ for details of these ports.)
iXvnc
Note that as of 3.3.3r2, Xvnc includes the functionality of iXvnc already. However
Andre's page is still the best documentation for how to use it at the moment.
Andre Moreira <andre@dei.isep.ipp.pt> has written some patches for Xvnc to allow it to
be run under the control of inetd and hence be started automatically as users connect. See
http://www.dei.isep.ipp.pt/~andre/extern/ixvnc.htm for details.
Share a Single Window
Céline FAGES <celine.fages@silogic.fr> has written some modifications for WinVNC to
allow it to share a single window.
See http://www.uk.research.att.com/vnc/contrib/one−window.txt for the modification
details.
See Céline's original message in:
http://www.uk.research.att.com/vnc/archives/2000−02/0658.html
Rudi De Vos has done further work based on Celine's patches. See
http://services.simac.be/vnc/single/index.htm (binaries available).
Wallpaper disabling version of WinVNC
Note that as of 3.3.3r8, there is a RemoveWallpaper option in standard WinVNC − see
the WinVNC documentation for details.
Jeremy Peaks jeremy@jerpeaks.freeserve.co.uk has amended WinVNC 3.3.3r7 to allow
you to automatically disable the wallpaper while someone is connected to your PC. This
cuts down on needless graphics being sent to the viewer, which can make VNC
significantly faster. See http://www.jerpeaks.clara.net for details.
Proxy for session capture/replay
Tim Waugh twaugh@redhat.com writes:
I have written two programs called rfbproxy and rfbplaymacro.
rfbproxy can be used for session capture and replay. Session replay can be looped, and
cycling between multiple capture sessions is supported. (I use it for doing demos at
shows.)
http://people.redhat.com/twaugh/rfbproxy/
rfbplaymacro can be used to provide scriptable control of VNC sessions. It uses a simple
scripting language to control the position of the mouse, keys to press, etc. It can be used
for quality assurance for graphical applications, scriptable control of otherwise
unscriptable programs, and so on.
http://people.redhat.com/twaugh/rfbplaymacro/
VNC Tight Encoder
Const Kaplinsky const@ce.cctpu.edu.ru writes:
VNC Tight Encoder adds low−bandwidth optimizations to VNC (both Unix and
Windows versions). Compression ratios achieved by "tight" encoder are usually 5−75%
higher than pure zlib compression. Additionally, this encoder is often much faster than
zlib at both server and client sides. Other enhancements like "local cursor" feature are
either implemented or in progress. Modified servers and viewers do not extend current
RFB protocol and thus are fully compatible with the standard VNC software. The Tight
Encoder homepage is at http://www.ce.cctpu.edu.ru/vnc/
x0rfbserver − accessing a standard X server with VNC
Jens Wagner has written some VNC−related tools, amongst which is a program called
x0rfbserver. This is a VNC server which serves a standard X server desktop thus
behaving more like WinVNC and MacVNC than does Xvnc. It is available in the
rfb−n.n.n.tar.gz package from http://www.hexonet.de/software.en/
Dual Monitors in Windows
Jay Freeman has done a patch for multiple monitor support in Windows 2000. See his
message at http://www.uk.research.att.com/vnc/archives/2001−02/0365.html
Policy Editor Template for Windows
Daniel Quinlan has written a policy template for VNC on Win32, which allows VNC
settings to be fixed across a whole domain. Useful in big domains! See
http://projects.chaosengine.net/winvnc/winvnc.ADM
Fastpush − Batch File Installer for Windows NT
Richard Harris has developed a batch file for Windows NT that let's it push VNC out and
install it as a service to other NT boxes. See http://www.darkage.co.uk/vnc/
IPv6 patch for X VNC
Mohacsi Janos mohacsi@ik.bme.hu writes:
Jun−ichiro itojun Itoh itojun@iijlab.net from the KAME project created an IPv6 patch for
VNC 3.3.3r2 available from
ftp://ftp.kame.net/pub/kame/misc/vnc−3.3.3r2−v6−20010421.diff.gz.
VNC Frequently Asked Questions (FAQ)
This isn't intended to be an introduction to VNC − have a look at the Getting Started page
first.
This list is being updated regularly, so, particularly if you are reading a local or mirrored
copy, it's worth checking the original from time to time. The archives of the mailing list are
also available, and they can be a good source of help − try searching them here.
Getting Started
1. Where can I get VNC?
2. I thought this was something to do with ORL? What's the Olivetti/Oracle link here?
3. I downloaded the .EXE files and they don't work!
4. I ran SETUP.EXE but it didn't work/nothing happens.
5. I ran WinVNC, but nothing seemed to happen. I don't see an icon on my taskbar.
6. WinVNC causes a 'Blue Screen of Death' on NT!
7. I get an error message about 'disabling Nagle's algorithm'. What does it mean?
8. I get an error message about 'Planar vs Chunky' screen formats. What does this mean?
9. How do I set up a Windows 95/98 machine so that I can dialinto it directly and view it using VNC?
10. I ran the Unix vncserver program and got a 'No such file or directory' message.
11. I ran the vncserver program and got a 'Number found where operator expected', 'prototype mismatch' or 'out
of memory' message.
12. I started the X server using vncserver, but it dies with a message "Could not open default font 'fixed'. "
13. I get errors like "failed to bind listener" and "Failed to establish all listening sockets" in the log file.
14. Vncserver seems to be dying quietly without putting any messages in the log file.
15. My Solaris/HP−UX Xvnc dies with a core dump!
16. I connected to my Unix VNC server and I just see a grey desktop with a cursor.
17. The log file is showing an error message from xrdb / Xlib.
18. The Java client doesn't work in my browser.
19. The Windows viewer dies as soon as it starts, or gives an error message (invalid page fault) immediately,
before I've even typed in a server name!
20. My viewer failed to connect to my server!
Running the programs
21. My X VNC server is working, but I don't see my normal environment. How can I change the Window
manager etc?
22. Can I remote the normal X display of my workstation (display :0) in the same way as the Windows server
does?
23. Why can I only run vncserver/Xvnc as root?
24. What X Visual does Xvnc use?
25. Can I cut and paste between the viewer and the server?
26. There's a memory leak in Xvnc!
27. Can I run the Windows server before anybody has logged in?
28. Why doesn't Ctrl−Alt−Del work? Why can't I unlock my NT workstation remotely? Why can't I stop the
screensaver remotely?
29. When I connect using VNC and then log into my Windows machine, I get disconnected and have to
reconnect!
30. Can I arrange for the local physical screen, keyboard mouse on my WinVNC server to be disabled when I'm
accessing it remotely?
31. I have troubles sending Ctrl−Alt−Del to a Windows server.
32. I can't make it work on Windows 2000
33. The dead keys don't work on my keyboard
34. When I start a DOS window, it doesn't display at the remote end.
35. I can't type into a DOS window or any DOS apps
36. The keyboard doesn't work / keys do strange things!
37. Most of my Windows apps work fine remotely, but this one doesn't update its window...
38. WinVNC is putting a huge load on my PC!
39. WinVNC dies, or causes other applications to die, after a short time when a viewer is connected.
40. My remote Windows display is appearing in a very garbled form.
41. I have problems with double−clicks in Windows.
42. I'm installing WinVNC as a service on lots of machines, and I get prompted for the password on each one.
Can I do this non−interactively?
43. I only have a two−button Windows mouse, and I really need three buttons for X...
44. I have a three−button Windows mouse, but the middle button doesn't work.
45. My machine doesn't have a DNS entry or static IP address! What address should other people use to
connect to my WinVNC server?
46. Can I get rid of the taskbar icon created by WinVNC?
47. Can I set up WinVNC to use my Windows NT password for authentication?
48. Can I make the Macintosh server start automatically when the machine boots up?
49. When I try to set a new password for my Mac server, I hit CHANGE on the web page, but nothing happens.
It is trying to bring up the settings.html page, but fails.
General Questions
50. How do I make VNC go faster?
51. Will VNC work through a firewall?
52. Which TCP/IP ports does VNC use?
53. Can I run VNC over a port normally used for a standard service? (eg. port 21, or port 80)
54. How secure is VNC?
55. Are you going to make it more secure?
56. Could you do file transfer (e.g. by drag drop) between the two ends?
57. Are you planning support for AIX, EPOC,HP−UX, SGI, Win 3.1, or my favourite platform ?
58. Would things work better if you compressed the stream?
59. Have you thought about caching bits of the screen at the viewer end?
60. Can I use VNC over a modem without using TCP/IP?
61. Does VNC have any Y2K (Year 2000) bugs?
62. How can I install WinVNC on multiple machines?
63. Can I connect multiple users to the same Windows server, and have them each see their own desktop, as
with WinFrame, NTrigue, WTS, etc?
64. Any other tips?
65. You misspelled 'organization' on the download page!
Compiling the source
66. I'm trying to compile WinVNC and the compiler complains about various missing files!
67. I'm having trouble compiling VNC on my platform...
VNC Frequently Asked Questions
Getting Started
Q1 Where can I get VNC?
The latest versions of VNC and this documentation are always available from the
ATTLabs Cambridge web site at http://www.uk.research.att.com/vnc. In addition,
various add−ons, extras, and ports to other platforms can be found on the contribs
page.
Q2 I thought this was something to do with ORL? What's the Olivetti/Oracle link here?
In January 1999, ATacquired ORL, the Olivetti Research Laboratory founded 12
years earlier, and recently jointly funded by Oracle, to create ATTLaboratories
Cambridge.
Q3 I downloaded the .EXE files and they don't work!
Netscape used to have a nasty habit of sticking a .EXE extension on a file when it
thinks it knows what type it is. When you fill in the download form you specify the
type of compression you would like to use, and this does not include .EXE, so if the
thing you downloaded has that extension, change it to whatever you requested
(usually .tgz) and then unpack the files using an appropriate decoder.
Q4 I ran SETUP.EXE but it didn't work/nothing happens.
Setup often doesn't work if you're running it from a network drive. Try copying it to
a local drive first. If you're on NT, make sure you have NT service pack 3 or later
installed. It appears that the InstallShield system we use won't work without it.
Q5 I ran WinVNC, but nothing seemed to happen. I don't see an icon on my taskbar.
If you are using a version earlier than 3.3.3, we suggest you upgrade. If you are
using 3.3.3 or later, this is because WinVNC doesn't know that somebody is logged
in. On Win95, this will happen if you bypass the 'password' prompt at login. If you
do this often, you might want to consider setting an empty Windows password, so
you don't get prompted!
If you're running WinVNC as a service, it may be that the service has not been
contacted by the 'Service Helper' when you logged in, to tell it who you are. (See the
WinVNC docs for more details.) You can try running the Service Helper from the
Start Menu to see if this is the problem.
Q6 WinVNC causes a 'Blue Screen of Death' on NT!
Again, make sure you have Service Pack 3 or later. If this happens it is usually due
to bugs in Service Pack 1. Also note the Service Packs' warning that if you add or
change components on your system you should reapply the service pack. If this
doesn't fix it, check for updates to your network and graphics drivers. VNC makes
very extensive use of both your video system and your network, and has a tendency
to find any bugs in either of them! There may well be bugs in WinVNC, but we
know of people running it on thousands of machines without problems, so please
check other parts of your system before assuming it's directly a WinVNC problem.
Q7 I get an error message about 'disabling Nagle's algorithm'. What does it mean?
Nagle's algorithm is an optimisation which is used most of the time on socket
connections. Under certain circumstances it batches up lots of small transmissions
into bigger ones, to make better use of the bandwidth. VNC relies on small
transmissions (like mouse movements) getting through quickly, so we turn it off
using the TCP_NODELAY option. If this causes an error, it is a problem with your
TCP/IP stack. In particular, Microsoft have broken this with NT service pack 6, but
we hear that the hotfix mentioned in KB article Q245678 will fix it. If this doesn't
work, you could try the security patch mentioned at
http://www.microsoft.com/technet/security/bulletin/ms99−046.asp. We've heard that
this fixes it as well, though this may just be a fix for an incomplete SP6a. In a future
version of VNC we will pop this up as a warning without making it fatal.
Q8 I get an error message about 'Planar vs Chunky' screen formats. What does this
mean?
This error usually occurs when running WinVNC on a Windows 95/98/ME or
Windows 2000 machine with a 16 colour (NOTE: NOT 16−bit) display mode. The
problem is due to the way the 16−colour VGA display driver represents the screen
internally. Upgrading the display drivers to hardware−specific ones or increasing the
number of colours (to 256 colours, 16−bit or 32−bit, for example) should fix the
problem. Windows NT does not exhibit this problem.
Q9 How do I set up a Windows 95/98 machine so that I can dial into it directly and view it
using VNC?
You need to install your modem from the control panel if you haven't already, and
you need to set up the dialup networking server on your server machine. (This is
included with Win98 and NT4. On Win95 it is in the Plus! pack, but you need to get
an update to version 1.3 or later from Microsoft's site. At the time of writing it can
be found here.) You can enable the dialup server from the 'Connections' menu of the
dial−up networking window. If it isn't there, or if you've updated the dialup
networking as mentioned above, you need to install it using the Windows Setup
section of 'Add/Remove Programs' in the control panel. When it's running and
you've dialled in, the server machine should have an IP address something like
192.168.55.1 as seen from the viewer − you can find this out by hovering over the
VNC server icon, using 'netstat −r', or running winipcfg. You should then be able to
connect to, for example, 192.168.55.1:0.
Q10 I ran the Unix vncserver program and got a 'No such file or directory' message.
Vncserver is a Perl script and so the first line has to point to the place where Perl is
installed. On Linux this will generally be /usr/bin/perl, on other platfroms it's
more likely to be /usr/local/bin/perl. Edit the line to point to the right
place and it should work.
Q11 I ran the vncserver program and got a 'Number found where operator expected',
'prototype mismatch' or 'out of memory' message.
Note: This problem should be fixed in 3.3.2R3 and later. This is a Perl error, and
happens on platforms where the Perl installation is not quite correct. We've heard of
it particularly on S.u.S.E. linux and Redhat 5.2. The problem is probably the line in
the vncserver script which says:
eval 'require "sys/socket.ph"';
Your first option is to change this to:
($] >= 5.0) ? eval 'use Socket' : eval 'require
"sys/socket.ph"';
If this doesn't work, particularly on S.u.S.E., you could try:
eval 'require "linux/socket.ph"';
or check your Perl installation to see if socket.ph is to be found in another directory.
Remember that vncserver is only a convenient wrapper around Xvnc. If it causes
you problems, feel free to discard it and run Xvnc directly.
Q12 I started the X server using vncserver, but it dies with a message "Could not open
default font 'fixed'. "
The 'fixed' font is needed for the server to start − if it can't find it, you need to
specify the correct font path for your machine in the 'vncserver' script. If you're not
sure what the path should be, type 'xset q' from within a normal X session. One of
the things reported is the font path used by your current X server, which is generally
the right thing to use for Xvnc. On some platforms you may need to use a colon as a
separator in the font path instead of a comma.
The VNC server can also get upset if you have directories on your font path which
don't actually exist on your system. Make sure you remove those. Also note that
older versions of the Xvnc server, by default, act as if they had a resolution of
100dpi. Some RedHat installations, for example, only install 75dpi fonts, so you
may need to install the missing font RPMs from your distribution or use the −dpi
option to Xvnc, or update to the latest version.
Q13 I get errors like "failed to bind listener" and "Failed to establish all listening
sockets" in the log file.
This is probably due to the permissions on /tmp/.X11−unix. You may well see this if
you update to Solaris 2.7 or Redhat 6.0, for example. See the section below entitled
"Why can I only run vncserver/Xvnc as root?".
Q14 Vncserver seems to be dying quietly without putting any messages in the log file.
Check that the Xvnc process really has died. If so, then check that your VNC font
path (set by uncommenting lines in the vncserver script) only includes directories
which actually exist. The XFree86 code in older versions of Xvnc seems to have a
problem which causes the server to die quietly if non−existent directories are
searched. Upgrading to 3.3.3 or later should reduce these problems.
Q15 My Solaris/HP−UX Xvnc dies with a core dump!
There is a bug in XFree86 (on which Xvnc is based) which makes Solaris servers
very unreliable if they have a pixel depth of 16. Use the −depth option to start your
server with a depth of 8 or 24, and you should be fine. We've heard that this can be a
problem with HP−UX as well. Darrell Kindred submitted a patch which speeds up
operation on Alpha machines, and also fixes this problem. The problem is also fixed
in 3.3.3 and later releases of VNC.
Q16 I connected to my Unix VNC server and I just see a grey desktop with a cursor.
After the vncserver script has started the Xvnc server, it then runs your
~/.vnc/xstartup script. By default this will try to start the twm window manager, but
if twm isn't on your path, or if you prefer something else, you can edit xstartup. The
log file may also give you clues about what is happening.
Q17 The log file is showing an error message from xrdb / Xlib.
By default, the first thing your xstartup script does is to run xrdb to load your
resources. So if the Xvnc server has not started for any reason, the xrdb is often the
first thing to notice it and print an error. (Though if you're getting a 'command not
found' message, then xrdb is probably not on your path − you need to find where it is
on your system and add it.)
If you get something like 'connection refused' or 'Can't connect: errno = 111', the
Xvnc server probably isn't there. So you should check whether the Xvnc process is
actually running, and whether there is anything earlier in the log file indicating why
it might have died. By far the most common reason for the server not starting is that
it can't find the 'fixed' font (see above). Other possibilities are that the server has
quietly crashed, or that it is taking a very long time to start up. The vncserver script
has a 3−second delay before running xstartup, but in extreme cases this may not be
enough. Lastly, the DISPLAY variable used by vncserver is based on the results
given by the 'uname − n' command. If your applications cannot resolve this to the
right IP address, perhaps because of funny settings in /etc/hosts, then they won't be
able to connect.
If you get a 'Client is not authorized to connect to Server' or similar message, there's
something wrong with the X authority setup − perhaps xauth is not on your path?
You could try using xhost to bypass this temporarily, but we wouldn't recomend this
as a long−term solution. There should be some indication in the log file if xauth has
failed.
Q18 The Java client doesn't work in my browser.
Several Java implementations have bugs which upset the VNC applet. Try pressing
Reload. We've seen this problem on Netscape Navigator 3 and Internet Explorer 5.
Note, too, that the Java applet often won't work if you are accessing the server via a
web proxy. Make a direct connection if you can.
Q19 The Windows viewer dies as soon as it starts, or gives an error message (invalid page
fault) immediately, before I've even typed in a server name!
There is a bug in the code which retrieves the list of most−recently accessed
connections for the initial dialog box. If, when you run the viewer for the first time,
you just hit return without entering a server name, this gets stored as an empty entry
in the 'most recently used' list. The fix is to use REGEDIT to delete the registry key
HKEY_CURRENT_USER\Software\ORL\VNCviewer\MRU. This will be fixed
in version 3.3.3r2 of the viewer.
Q20 My viewer failed to connect to my server!
VNC relies on a correctly−configured and operational TCP/IP network, so please
make 100% sure that your TCP/IP setup is right before you start asking questions on
the mailing list. Here are some things you should check before assuming it's a VNC
problem; consult your local expert if you don't know how to check them:
♦ Can you ping the server machine from the client?
♦ Is the VNC server definitely running on the server machine?
♦ The server listens on port 5900+displaynumber. Can you telnet to this port
from the client machine?
♦ Have you specified the address correctly to the viewer? If you're using a
display number other than zero for the server, (usually the case on Unix
machines), did you remember to specify it?
♦ Is the server name known to the DNS? Try using an explicit IP address
instead of a name (eg. 123.456.78.9:0 instead of snoopy:0).
♦ Do you have any firewalls or proxies in the way that could be blocking
acess?
♦ If using the java client, did you remember to specify the correct port as part
of the URL? (eg. http://snoopy:5800)?
♦ Can you try running either the server, or the client, or both, on different
machines on your network to find whether the problem is at one end or the
other?
♦ Can you try running the software on a different architecture? eg., if you are
having problems viewing a PC from another PC, can you try connecting
from a Unix machine?
Running the programs
Q21 My X VNC server is working, but I don't see my normal environment. How can I
change the Window manager etc?
The window manager is started by the ~/.vnc/xstartup script. We use twm, as this is
available on almost all Unix platforms. Edit the script if you'd rather replace it with
something else. On many platforms you can, as an alternative, just make xstartup a
link to whatever script normally starts your X environment.
If you want to be more sophisticated, you can specify the −name option to vncserver,
and then take different actions in the xstartup script based on the name given. For
example:
case "$VNCDESKTOP" in
kde)
startkde
;;
*)
xterm −geometry 40x10+40+40 −ls −title "$VNCDESKTOP Desktop"
twm
;;
esac
See also the next question.
Q22 Can I remote the normal X display of my workstation (display :0) in the same way as
the Windows server does?
It is possible to add VNC support to a standard X server, but we have no plans to do
so at the moment. We've been told that Jens Wagner has done this − see x0rfbserver.
We tend to run all our X sessions as VNC sessions and only use the local X server to
run the viewer. It's very fast when on the same machine as the server! If you feel that
it's overkill to run two X servers on the same machine, you might consider Ganesh
Varadarajan's svgalib−based viewer.
Q23 Why can I only run vncserver/Xvnc as root?
The most likely reason for this is that Xvnc can't create the unix domain socket (the
path for this unix domain socket is usually /tmp/.X11−unix/Xn). Try making sure
that users can write to this directory by making it world−writable, i.e. chmod
01777 /tmp/.X11−unix
You can also tell Xvnc not to use unix domain sockets by giving it a −nolisten
local option − note that this means DISPLAY must be set to "host:n", not just
":n".
An alternative is to set the Xvnc binary to have the same permissions as your normal
X server, but this may be more of a security risk.
Q24 What X Visual does Xvnc use?
By default, vncserver will start Xvnc with the same depth as the current X display, if
there is one, or 8 bits deep if there isn't. We've tried to steer clear of colour maps as
much as possible and normally use "true colour", even when there are only 8 bits per
pixel.
Unfortunately some X applications don't cope too well with an 8 bit TrueColor
visual. You can make Xvnc use the more normal PseudoColor visual by giving a
"−cc 3" option to vncserver.
Q25 Can I cut and paste between the viewer and the server?
VNC supports copying and pasting of ASCII text in both directions, provided the
viewer and server allow it. When the clipboard changes on the machine running the
viewer, the changes are copied to the server and vice versa. Some notable
exceptions:
♦ X has more than one method of using the clipboard and different
applications do it different ways. Emacs and xterm should just work. If you
find that your X application doesn't work via VNC, you can generally use
the xcutsel program to copy the clipboard between the different X
methods. VNC uses Cut_Buffer0, so if you select text in Unix Netscape, for
example, you may need to click 'Copy PRIMARY to 0' before it is
accessible at the other end of the VNC link. You can use X resources to
make the button labels more meaningful. For example, here's a script:
#!/bin/sh
exec xcutsel
−xrm
−xrm
−xrm
−xrm
−xrm
−xrm
\
'*quit.borderWidth:0' \
'*quit.height: 1' \
'*quit.label:' \
'*sel−cut.label: Clipboard: out of netscape' \
'*cut−sel.label: Clipboard: into netscape' \
'*font: −*−helvetica−*−r−*−*−*−*−*−*−*−*−*−*'
Michael Witrant has written a program to do the transfer automatically. He
writes:
I'm glad to announce autocutsel version 0.1.
People using xcutsel to copy/cut and paste between VNC and an X desktop
might be interested with it. I was bored clicking on xcutsel's buttons to
copy/paste between GTK apps on my VNC desktop and the Windows
system running vncviewer.
This tool regularly scans the primary selection and the cutbuffer 0. If one of
them is changed, it updates the other one.
I don't need xcutsel anymore and have a working cut and paste between
GTK (through VNC) and Windows.
You can get it there: http://www.lepton.fr/tools/autocutsel
♦ Java applets running in the browser cannot access the clipboard of the
machine on which they are running, so the Java viewer has a clipboard
button. This pops up a window displaying the contents of the remote
clipboard, which should allow you to manipulate it locally.
Q26 There's a memory leak in Xvnc!
This is fixed in versions 3.3.2r3 and later. If you're using an older distribution you
can find a patch for it here.
Q27 Can I run the Windows server before anybody has logged in?
Yes. Read the section on running WinVNC as a service in the documentation.
Q28 Why doesn't Ctrl−Alt−Del work? Why can't I unlock my NT workstation remotely?
Why can't I stop the screensaver remotely?
Make sure you are running a recent version of VNC, and that you are running it as
a service. From some platforms you will not be able to type Ctrl−Alt−Del directly,
because it will be caught by the local machine. The Windows viewer, for example,
has an option on its menu to send a Ctrl−Alt−Del to the remote host. In some
situations, you will find that something like Ctrl−Alt−Backspace or Ctrl−Alt− may
work instead. Screensavers sometimes use a different resolution and so can
disconnect you when they stop or start − see the next question.
Q29 When I connect using VNC and then log into my Windows machine, I get
disconnected and have to reconnect!
Sometimes logging in will involve a change in screen resolution, if the user's display
settings are different from the defaults. If this happens, the server will disconnect
you and you will need to reconnect to get the new screen size. Just occasionally on
NT, the mode seems to change resolution temporarily as you log in, and if WinVNC
happens to see this you can also be disconnected, even if the final resolution
changes.
If the user has set a different display number in their personal WinVNC properties
dialog you will also be disconnected.
Q30 Can I arrange for the local physical screen, keyboard mouse on my WinVNC server
to be disabled when I'm accessing it remotely?
We don't know of a way to do this which will still allow us to read the screen and
inject fake input events from WinVNC. If you have any code which might help, we'd
be pleased to hear from you! Until then, if this aspect of security is important, we
can only recommend that you consider other options such as password−protected
screen savers, lockable keyboards, turning off the screen, unplugging the keyboard,
or other physical security devices.
Q31 I have troubles sending Ctrl−Alt−Del to a Windows server.
◊ Ctrl−Alt−Del will only be recognised by a Windows NT VNC server, and
only when VNC is running as a service. On Win95/98, it suspends all
processes, including the VNC server, so it wouldn't be much use!
◊ The Windows, X and Java viewers have menu options to send Ctrl−Alt−Del,
so they don't get interpreted locally. The X viewer's menu is usually
accessed by pressing F8.
◊ If you are using the Mac or X−based viewer, you should just be able to type
the keys, but some platforms seem to catch the keystrokes. If this is the case
for you, I'm afraid you'll have to work out how to stop it! It might be worth
trying slight variations (eg. the right Ctl key, or a different Del/Delete key, if
you have one).
Q32 I can't make it work on Windows 2000
If you're using the standard VGA driver, can you change to SVGA or something
specific to your video card? Windows 2000 uses the old Win98 VGA driver, which
has problems.
Q33 The dead keys don't work on my keyboard
It's quite a challenge working out how to do international keyboard support across
different platforms. We'll sort out dead keys, but if you'd like to work out how to do
them, we'd be grateful.
Q34 When I start a DOS window, it doesn't display at the remote end.
You might also notice the pointer leaving a trail of arrows behind it. This happens
when your DOS sessions are full−screen. WinVNC cannot read the display when
this happens, but keystrokes should still get through. Press Alt−Enter to switch the
DOS box to windowed mode.
Q35 I can't type into a DOS window or any DOS apps
This derives from the fact that Windows 95 uses BIOS calls and not Windows
messages to get keystrokes within the command prompt. It should be fixed in
version 3.3.1 R19 and later of the WinVNC server, so if you're using an earlier
version, uninstall and upgrade.
Q36 The keyboard doesn't work / keys do strange things!
There is one common problem which can cause this. If a modifier key, such as Shift,
Ctrl or Alt, is pressed, and the viewer window then loses focus or dies, the 'key
release' message never gets to the viewer and hence never gets to the remote server.
The remote machine will then think that M is Ctrl−M etc. We have done various
things to reduce the chance of this happening; the viewers release various modifiers
automatically when they lose focus, for example, but it can still occur and can be
confusing when it does. The solution is easy: simply press and release the modifier
key which is stuck. If you don't know which it is, then try them one at a time.
Q37 Most of my Windows apps work fine remotely, but this one doesn't update its
window...
WinVNC can use a variety of hints to guess when a particular area of the screen has
changed. The most useful is the occurrence of a WM_PAINT message in an
application. Not all applications use these messages, though; the Windows clock is a
good example.
In the registry WinVNC has a list of application names, and which things to use as
update clues. A few standard applications are set up by the 'Install default registry
settings' in the WinVNC section of the Start Menu; it tells WinVNC to look for
timer messages from the clock, for example. Make sure you install these if you
haven't already. Any other apps you run will appear in the registry, so you can easily
tweak the settings. For more info, see the WinVNC documentation.
Q38 WinVNC is putting a huge load on my PC!
Firstly, WinVNC should have negligible impact when nobody is connected to it. It
does practically nothing in this state. If you find that it is running at something close
to 100% CPU when there is a remote connection, check the Update Handling in the
Properties dialog box. The default settings should be fine in most circumstances. If
you have either (a) Ticked 'Poll Full Screen' or (b) Ticked 'Poll Foreground Window'
or 'Poll Window Under Cursor' without having 'Poll Console Windows Only' ticked,
then the load will be much higher. See also the section below on 'How do I make
VNC go faster?'.
Q39 WinVNC dies, or causes other applications to die, after a short time when a viewer is
connected.
Some screen savers, particularly on Windows 95, do not interoperate well with
WinVNC. Try disabling all screen savers on the machine running the server and see
if that fixes it.
Q40 My remote Windows display is appearing in a very garbled form.
There can be problems with WinVNC being unable to detect changes to the screen
on certain applications, as mentioned in the previous question, but there should not
generally be serious screen corruption.
If you have Win95/98 and you have a video driver which only uses 16 colours, you
need to change it. VNC really expects at least 256 colours, though 16 shoudl work
on NT.
Make sure you have the latest service packs installed for your OS and experiment
with different video drivers. WinVNC relies on various aspects of the video card and
driver; in particular its ability to BitBlt correctly, and several users have solved their
problems by updating the driver.
Q41 I have problems with double−clicks in Windows.
By default, mouse button up events are used to signal to WinVNC that the current
window may have changed and should be re−scanned. This can introduce a delay
which results in a double−click being interpreted by the system as a pair of single
clicks. Variations in network delays can also cause the clicks to be too widely
separated when they appear at the server. Some suggestions:
◊ In many situations, the action performed by a double−click is also available
from a right−button menu.
◊ You may find that clicking three times, rather than two, will help.
◊ You can change the double−click speed using the control panel on the server
so that more widely−spaced clicks are still interpreted as doubles.
◊ You can disable the left−click hook by editing the registry, as described in
the WinVNC documentation, although this may mean that WinVNC misses
some screen updates from that application.
Q42 I'm installing WinVNC as a service on lots of machines, and I get prompted for the
password on each one. Can I do this non−interactively?
Not directly, but you can by editing the registry before installing. On a machine
which already has WinVNC installed, copy all the registry settings under
HKEY_USERS/.Default/Software/ORL/WinVNC3
into the same place in the registry of the remote machine, either by hand or using
your favourite tool. When WinVNC runs, it will see the password in that section of
the registry and will not prompt you to type one in.
Another hint that might be useful was sent in by Johannes Norinder. If you run the
SETUP program with a −r option, it will create a setup.iss file, probably in your
%systemroot% directory. Following installations on similar computer can be started
with "setup −s −h1[path to your setup.iss]" and you won't have to answer the
questions again!
Q43 I only have a two−button Windows mouse, and I really need three buttons for X...
Get a recent version of the Windows viewer. This allows you to emulate the middle
button by pressing both buttons together.
Q44 I have a three−button Windows mouse, but the middle button doesn't work.
This is almost certainly a problem with your mouse driver. The Windows VNC
viewer recognises standard WM_MBUTTON... messages and should work with any
driver that generates them. Some drivers, knowing that Windows seldom normally
uses the middle button, either don't recognise it, or map it to something else like a
double−click. Try telling Windows that you have a different type of mouse, (we've
had reports that the Logitech PS/2 Port Mouse is a good one to try) or use the
3−button emulation mode mentioned above.
Q45 My machine doesn't have a DNS entry or static IP address! What address should
other people use to connect to my WinVNC server?
If your ISP allocates you a dynamic IP address when you dial in, you will need to
give that to the person trying to connect. On Windows95 machines, after connecting,
you can use the winipcfg program (type it into the Start/Run... box). On NT
machines, type ipconfig at a command prompt. On linux machines, try
hostname −i. A handy new feature on recent versions of WinVNC causes the IP
addresses of the local machine to be displayed when the mouse hovers over the
WinVNC icon, (if they can be determined at that time).
Of course, if your phone line is now in use you may need to find some other way of
getting this information to the remote person. I suggest a chat system like AOL
Instant Messenger. There are also programs out there which will automatically
create a web page with your current IP address, which the other person could then
read.
Q46 Can I get rid of the taskbar icon created by WinVNC?
No. Not without changing and recompiling the source code. We feel that there would
be few legitimate uses of VNC where you would need to conceal its operation.
However, there might be occasions where you would not want the user of the server
machine to be able to adjust the WinVNC settings, password, etc, so versions later
than 3.3.2R3 include a 'restricted' mode where the icon is still visible, but the menu
options normally available from it will be disabled. See WinVNC − Advanced
Settings in WinVNC docs for the details.
In some situations, particularly on Win95, you may find that the icon is not
displayed. This is not a feature, it's a bug which we plan to iron out! See the question
above on 'I ran WinVNC don't see an icon in my Taskbar!'.
Q47 Can I set up WinVNC to use my Windows NT password for authentication?
Not at present. Partly because there are many problems with the NT security model,
but chiefly because we want to keep VNC as cross−platform as possible. In the
future we may try and make the code and the protocol more modular so it will be
easier to add your own favourite authentication.
Q48 Can I make the Macintosh server start automatically when the machine boots up?
Yes, create an alias to VNCServer and put it into the Startup items folder (in your
system folder).
Q49 When I try to set a new password for my Mac server, I hit CHANGE on the web page,
but nothing happens. It is trying to bring up the settings.html page, but fails.
The http portion of the current server is unstable, sorry. But you can set your settings
with AppleScript.
Open the AppleScript editor and enter the following:
tell application "VNCServer"
set password to "whatever"
end tell
There are other options you can set, which are all documented in the directory inside
VNCServer (use the AppleScript editor to open the dictionary)
General Questions
Q50 How do I make VNC go faster?
We find VNC to be perfectly acceptable as our normal method of accessing Unix
desktops on a daily basis. This is over a 10 Mbit/s ethernet on reasonably modern
machines, using the X or Win32 viewer. Because Windows gives us fewer hints
about what it's doing, and because we don't have the source code for Windows in the
same way that we do for X, the WinVNC server has to work harder to find out
what's changed, and so a really fast machine should make a big speed difference. For
more information about how the Windows server works, see the WinVNC
documentation. But if you've been disappointed by the speed of the Windows server,
don't give up. We're improving it gradually, but it'll be a while before it's as fast as
on Unix.
There are several things that can slow any VNC session down, however, and you
may like to consider these if you find it too slow:
♦ Unusually 'busy' desktops. The VNC protocol is very efficient at rendering
areas of a single colour, such as you generally find on window title bars,
scrollbars, backgrounds of pages etc. But if, for example, you have pretty
24−bit photographs of your girlfriend as your screen background, or dithered
title−bars on your windows, you may pay a price for the aesthetics. A
colourful or patterned desktop background will probably slow down
VNC more than any other single factor. We have some suggestions on
speeding up the twm window manager, some of which will also apply to
other environments.
♦ Hi−colour desktops. Don't use 24−bit colour if you can use 16 or 8 equally
well. Remember, on Unix you can run multiple servers, so I have a big
16−bit desktop for normal work and a small 8−bit one for when I log in
from home. The server can send out a wide range of pixel formats, and some
viewers will allow you to request a specific format for that session. On the
Windows viewer, for example, if you click Options... when making the
connection, you can request only 8−bit pixels from the server − useful if the
network gets slow. If you are using a modem, I recommend changing the
shortcut in the Start menu to include the /8bit option − this will then be the
default. Similarly, if you regularly connect to a remote WinVNC server,
consider whether you could run happily at lower resolution. A 1280x1024
screen has more then 4 times as many pixels as a 640x480 one, and if all you
are doing is checking a printer queue you probably don't need them all!
Note, though, that on WinVNC, 16−bit colour is usually the best to use. See
below.
♦ Elderly graphics cards or drivers may make quite a difference; this is a
graphics−intensive application! On Windows the graphics system on the
server will affect the speed as well as the one on the viewer.
♦ Some applications are not very economical about redrawing their display.
Early versions of Unix Netscape, for example, tended to draw everything
twice when scrolling, which did nothing to help the smoothness under VNC.
X11Amp flashes its display very fast when in 'pause' mode.
♦ Some Java Virtual Machines are particularly fast at reading from the
network and particularly slow at drawing to the screen, or vice versa. With
the Java viewer it is worth experimenting with the encodings available from
the Options menu, as we sometimes find big differences in speed.
♦ If you are connecting to WinVNC, don't change the default settings in the
Properties box unless you need to. See 'WinVNC is putting a huge load on
my PC' above.
♦ All the standard clients can do local rectangle copying, and this generally
means that dragging a full window on X is much faster than dragging its
outline. If your window manager allows this, it may speed things up. Roman
Mitnitski also reports that when he increased the mouse dragging threshold
(−t option of the server) the performance was improved.
♦ Generally, with WinVNC, use 16−bit colour (65536 colours) on the server if
you can. 16−bit is almost always the best depth to use, because:
1. 256−colour screens have to be palette−converted before they can be
transmitted to truecolour clients. Only if the client is 256−colour
palette− based will you see any performance increase. Even if the
client is 256 colour truecolour, it'll have to convert via a 32−bit
truecolour palette!
2. 24−bit screens have to be specially munged via 32−bit since VNC's
internal colour−handling routines don't work with 24−bit directly.
3. 24 and 32−bit screens have to have each pixel looked up in three
tables to get the converted value.
4. Graphics cards claiming to do 24−bit often actually do 32−bit with
munging − this in many cases makes 24−bit slower just for general
use than 32−bit!
5. Finally, 16−bit involves no palette processing and a single lookup in
a cached src_format to dest_format table to convert the pixels.
On slow links, you may also want to use some software to compress the data
between the two ends. SSH or Zebedee can be used for this. See the section on
security below.
Q51 Will VNC work through a firewall?
It depends on your firewall, and whether you want to access a server inside your
firewall from elsewhere, or a server outside your firewall from inside.
Generally firewalls are designed to prevent incoming connections except to certain
well−known machines and ports. If you can configure these to include your VNC
server, then you will be able to access it from anywhere in the world. There is a good
argument to be made for the fact that VNC is less of a security risk than X, so if your
site doesn't allow X in or out it may still allow VNC.
Many modern firewalls will allow outgoing connections initiated from inside, so you
can often access servers on outside machines. It is straightforward, for example, to
recompile the viewer source to include SOCKS support, or to make other special
arrangements. See the contribs page.
It's a pity that Java within a browser doesn't automatically use SOCKS if the browser
is configured to use it. There's probably Java SOCKS support out there somewhere...
If your internet access is through a router which does Network Address Translation,
you may be able to configure the router to redirect particular incoming ports to
particular machines. So you could run WinVNC with a display number of 0 on
machine snoopy, and with display 1 on machine woodstock, then set your router to
send port 5900 to snoopy and 5901 to woodstock. See below for information on the
other port numbers used by VNC.
Q52 Which TCP/IP ports does VNC use?
A VNC server listens on two ports. The exact port numbers depend on the VNC
display number, because a single machine may run multiple servers. The most
important one is 59xx, where xx is the display number. The VNC protocol itself runs
over this port. So for most PC servers, the port will be 5900, because they use
display 0 by default.
In addition, VNC servers normally have a small and very restricted web server built
in, which allows you to connect a browser to them and use the Java viewer. This
runs on port 58xx. Note that this is the HTTP port used for downloading pages and
applets, but once the applet is running it uses 59xx for VNC just like any other
viewer.
The servers can be changed to listen on other ports if, for any reason, these are not
suitable for you. See the server's documentation for more details. Most of the
viewers, if given a display number larger than 99, will interpret it as a direct port
number and will not add 5900. See also the next question.
If you are running a viewer in 'listening' mode, where it accepts connections initiated
by the server, it will listen for incoming VNC on port 5500.
Q53 Can I run VNC over a port normally used for a standard service? (eg. port 21, or port
80)
In rare circumstances, people may want to do this, perhaps because they have a
firewall which only allows connections to certain ports. This can be done, at least for
the Windows and Unix servers (see their documentation), but the following points
need to be borne in mind:
◊ On some systems (eg. most forms of Unix), ordinary users are not allowed
to run servers on ports below 1024.
◊ You obviously can't run a VNC server on a port that's already being used for
other things.
◊ Many VNC servers use two ports: one for the VNC server, and one for the
HTTP server that provides the Java applet (see previous question). If you
plan to use the Java viewer, you may want to change both. Not all servers
will allow this at present.
◊ You need to tell the viewer the right display number. Normally, display
numbers come between 0 and 99. If you specify any number smaller than
99, the viewers add 5900 to get the port number. If you specify a larger
number, the viewers take it as a port number directly. So how do you use
port numbers lower than 99? You have to specify a negative display
number! For example, to connect to a server running on port 80 on machine
'snoopy':
vncviewer snoopy:−5820
because −5820 + 5900 = 80. This may not work with all viewers, but Unix
and Windows seem to be fine.
Q54 How secure is VNC?
Access to your VNC desktop generally allows access to your whole environment, so
security is obviously important. VNC uses a challenge−response password scheme
to make the initial connection: the server sends a random series of bytes, which are
encrypted using the password typed in, and then returned to the server, which checks
them against the 'right' answer. After that the data is unencrypted and could, in
theory, be watched by other malicious users, though it's a bit harder to snoop a VNC
session than, say, a telnet, rlogin, or X session. Since VNC runs over a simple single
TCP/IP socket, it is easy to add support for SSL or some other encryption scheme if
this is important to you, or to tunnel it through something like SSH or Zebedee.
SSH allows you to redirect remote TCP/IP ports so that all traffic is strongly
encrypted, and this can be combined with VNC. SSH can also compress the
encrypted data − this can be very useful if using VNC over slow links. See the
'Using SSH with VNC' page. Zebedee is a similar system which can be sometimes
simpler to use. You can find info here.
While we're on the subject of security, you should also be aware that only the first 8
characters of VNC passwords are significant. This is because the 'getpass' call used
in the Unix server to read a password has this restriction, and the other platforms
have been made compatible with this.
Wolfram Gloger < wmglo@dent.med.uni−muenchen.de> has built Xvnc with the
TCP Wrapper library, allowing you more control over which hosts are allowed to
connect. See the contribs page for details.
Q55 Are you going to make it more secure?
We do hope eventually to add better security to VNC, but there's also a good
argument for not doing so. If security is a concern, it can be better to use a single
system such as SSH, FreeS/WAN, or Zebedee to encrypt all your traffic, rather than
relying on the individual packages to do the right thing. Then, if you decide in a
year's time that one system is too easily crackable, you can replace it yourself and all
of your communications will benefit. It may also be easier to fit in with corporate
security systems this way.
Q56 Could you do file transfer (e.g. by drag drop) between the two ends?
It's certainly possible, but we aren't going to do it. File transfer seems simple, but is
actually rather complicated to do in a cross− platform way. Some examples: How do
you map filenames between platforms with wildly different naming conventions?
How do you cope with running out of disk space? What about permissions? How do
you do it on platforms which don't have a concept of drag and drop? Can you
interrupt the transfer if it's taking too long? Can you restart it if it gets interrupted?
Since there are a very large number of perfectly good systems out there for
transferring files over IP, we aren't planning to incorporate it in VNC. It would
probably more than double the size of the code, and would introduce all sorts of
issues that we, quite frankly, aren't interested in! So use the standard file transfer
built into Windows, or use FTP, or netcat, or the web, or rcp, or ssh, or...
Q57 Are you planning support for AIX, EPOC,HP−UX, SGI, Win 3.1, or my favourite
platform ?
We have provided VNC on all the platforms we use here, and it's difficult to provide
binaries for anything we don't have, and it takes a while to get up to speed on new
platforms. Information about third−party ports of VNC to a large number of other
platforms can be found on the 'contribs' page, so check there first. Remember that a
viewer is available for any platform which runs Java, though the speed may vary
quite a bit. But for many platforms it should not be difficult to compile at least the
viewer. If anyone tailors the sources for a particular platform we will happily either
incorporate the changes in the main source distributions or make the patches
available from our site.
Q58 Would things work better if you compressed the stream?
VNC incorporates really quite efficient compression in the sense that we generally
send a tiny fraction of the raw data, probably something like 1/20 on average. The
details are in the protocol spec if anyone's interested. On a couple of test screen
dumps we found that the Hextile encoding was more efficient than GIF! I don't
know whether this is true in general.
But we haven't done more general encoding after that; we've tended to the view that
(a) it might introduce too much latency and (b) most modems compress pretty well
anyway. We are planning some zlib−compression experiments in the near future to
see how this affects things. Because different bits of the screen can be sent using
different encodings, the server could, in theory, detect that one bit would be most
efficiently sent as JPEG, while another would be better hextiled.
The question is always how much work it's worth doing at the server to find this out.
To some degree you can control this already, because the viewers allow you to
specify your preferred encoding. Under X, if your viewer and server are on the same
machine the viewer will use the raw encoding by default, otherwise it will use
hextile. You may find that by selecting different encodings on the command line you
get better performance.
See also the suggestions above about using ssh or zebedee, which also provide
compression.
Q59 Have you thought about caching bits of the screen at the viewer end?
Yes, that could also be good. You could have an off−screen cache in the viewer and
the server could copy things from there to the screen. Management of this would add
a certain amount of complexity, though.
Since there is already a CopyRect primitive in VNC, an alternative approach would
be to copy updates from another part of the screen if they already exist there, rather
than resending them. Again, to make the server find out efficiently when this is
worth doing would be an interesting challenge, and volunteers for the project are
welcome!
Q60 Can I use VNC over a modem without using TCP/IP?
Not at present. VNC could run over other transports such as RS232, firewire, USB,
modems, IrDA etc, in fact, anything which gives a reliable 2−way connection. At
present we just use TCP/IP, because it's convenient, ubiquitous, and easy to route.
This means that you can use VNC over anything which supports TCP/IP, so using it
over a modem is just the same as any other network, once you have Dial−Up
Networking set up. If you need to communicate directly between two machines
without going via the internet/intranet, then set up a remote access server on one and
dial in from the other.
Q61 Does VNC have any Y2K (Year 2000) bugs?
The simplest answer to this, now that the date has passed, is that lots of people are
happily running it without any problems. However, if you want more details, this is
what we used to say here in the last millennium:
The WinVNC server and Windows viewer have been tested on a PC
with its date running through the 2000 boundary without any
problems, so unless the underlying OS or BIOS has difficulties,
VNC on a PC should be fine. The VNC part of the X−based Unix
VNC server only uses dates when writing the log files; the logfile
entries are timestamped with a two−digit year, but the format is easy
to change if required and the entries are not intended to be
machine−readable. The developers of the XFree86 server on which
Xvnc is based state that there are no Y2K problems (see
http://www.xfree86.org/FAQ/). We therefore issue the standard
disclaimer: we believe the VNC code, in its entirety, to be free from
Year 2000 problems, subject to the other components of the systems
on which it is running.
Q62 How can I install WinVNC on multiple machines?
When you run WinVNC for the first time on a machine, it will prompt you for a
password. If you are doing this on a large number of machines, especially remotely,
this can be a nuisance. The way to bypass this is to make sure that the target machine
already knows the password by putting the encrypted version straight into the
registry before installing. So:
◊ Install WinVNC on a master machine and set the password
◊ Copy the appropriate registry entries from
HKEY_CURRENT_USER\Software\ORL\WinVNC3 , or
HKEY_USERS\.DEFAULT\Software\ORL\WinVNC3,
(if you installed WinVNC as a service) and install them on the target
machine using your favourite registry utility.
◊ Copy the VNC files (typically under C:\Program Files\ORL\VNC) onto the
remote machine.
◊ Install as a service or into the startup directory as appropriate.
Q63 Can I connect multiple users to the same Windows server, and have them each see
their own desktop, as with WinFrame, NTrigue, WTS, etc?
No. Windows NT has a reasonable concept of multi−user access, but not where the
GUI is concerned. Basically, you can't do this without access to the source code of
Windows, and that's rather difficult to get hold of! We have successfully run
multiple WinVNC servers on a Windows Terminal Server machine, but they don't
update unless a WTS client is also connected, which rather defeats the purpose. It
may be possible in future to get documentation on the WTS APIs, in which case we
might be able to do something better.
Of course, if your server is Unix−based, then you have no problem. You can run
dozens of VNC servers on a single machine.
Q64 Any other tips?
Several people have indicated that they have to use Windows occasionally but prefer
to use Unix most of the time, and so want to access a PC under the desk from the
Unix box.
Here's a suggestion: all other things being equal, I recommend using the Windows
box to view the Unix machine rather than the other way around. This is chiefly
because Windows generally works better as a client than as a server, and also
because PC graphics cards are often better than those in Unix workstations.
Remember, you can create a VNC session of any size and pixel depth you like.
If you're very anti−Windows you can make your VNC desktop the same size as the
local screen and set the taskbar to 'Auto hide' and just pretend you're on an X
terminal, but pop up the Start menu when you have to use PowerPoint.... The
Windows viewer also now has a proper 'full−screen mode', so you don't even need to
bother with auto−hide.
Q65 You misspelled 'organization' on the download page!
No we didn't. We spell it like that in the UK. Actually, we spell it both ways, but the
's' spelling is more common, despite what the OED says! Now, as for 'misspelt'...
Compiling the source
Q66 I'm trying to compile WinVNC and the compiler complains about various missing
files!
You need to compile using the No_CORBA configuration, or it will try to include
various files which are part of our internal version only. If you still get errors about
vncControl.idl after doing this, blame Microsoft! It's a bug in the way Visual C++ 6
imports projects that were built under Visual C++ 5. You don't need the IDL file,
and it isn't included. Just delete it from the project.
Q67 I'm having trouble compiling VNC on my platform...
Have you checked the contribs page? Several people have provided hints on how to
build VNC on other platforms. If yours is not listed there, you might at least get
some clues..
Why are we giving it away?
We have been using thin−client systems at our lab for several years and have found
them very useful. We hope that there might be others out there who will find VNC
useful, port it, find any bugs in it, and give us feedback.
When we made our high−performance CORBA implementation, omniORB, publicly
available, it was exceedingly popular and so we are releasing VNC on the same terms
and conditions: those of the GNU General Public License.
We believe that systems based on stateless endpoints have many advantages,
particularly for users who frequently move about within one building or the local area,
because of the seamless mobility which comes from maintaining all the state at the
server. And with our Java client, these mobile workers can access their email etc
from anywhere in the world. But VNC is also useful for the enthusiast at home who has
a Windows and a Linux machine but only one monitor.
The release contains this documentation, binaries for several platforms, and full source
code. It is distributed in the hope that it may be useful, but comes without any
warranty, explicit or implied.
The GNU General Public License
VNC is distributed under the terms of the GNU General Public License as below.
Table of Contents
• GNU GENERAL PUBLIC LICENSE
♦ Preamble
♦ TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND
MODIFICATION
♦ How to Apply These Terms to Your New Programs
GNU GENERAL PUBLIC LICENSE
Version 2, June 1991
Copyright (C) 1989, 1991 Free Software Foundation, Inc. 59 Temple Place − Suite 330,
Boston, MA 02111−1307, USA
Everyone is permitted to copy and distribute verbatim copies of this license document,
but changing it is not allowed.
Preamble
The licenses for most software are designed to take away your freedom to share and
change it. By contrast, the GNU General Public License is intended to guarantee your
freedom to share and change free software−−to make sure the software is free for all its
users. This General Public License applies to most of the Free Software Foundation's
software and to any other program whose authors commit to using it. (Some other Free
Software Foundation software is covered by the GNU Library General Public License
instead.) You can apply it to your programs, too.
When we speak of free software, we are referring to freedom, not price. Our General
Public Licenses are designed to make sure that you have the freedom to distribute copies
of free software (and charge for this service if you wish), that you receive source code or
can get it if you want it, that you can change the software or use pieces of it in new free
programs; and that you know you can do these things.
To protect your rights, we need to make restrictions that forbid anyone to deny you these
rights or to ask you to surrender the rights. These restrictions translate to certain
responsibilities for you if you distribute copies of the software, or if you modify it.
For example, if you distribute copies of such a program, whether gratis or for a fee, you
must give the recipients all the rights that you have. You must make sure that they, too,
receive or can get the source code. And you must show them these terms so they know
their rights.
We protect your rights with two steps: (1) copyright the software, and (2) offer you this
license which gives you legal permission to copy, distribute and/or modify the
software.
Also, for each author's protection and ours, we want to make certain that everyone
understands that there is no warranty for this free software. If the software is modified by
someone else and passed on, we want its recipients to know that what they have is not
the original, so that any problems introduced by others will not reflect on the original
authors' reputations.
Finally, any free program is threatened constantly by software patents. We wish to avoid
the danger that redistributors of a free program will individually obtain patent licenses,
in effect making the program proprietary. To prevent this, we have made it clear that any
patent must be licensed for everyone's free use or not licensed at all.
The precise terms and conditions for copying, distribution and modification follow.
TERMS AND CONDITIONS FOR COPYING,
DISTRIBUTION AND MODIFICATION
0. This License applies to any program or other work which contains a notice placed by
the copyright holder saying it may be distributed under the terms of this General Public
License. The "Program", below, refers to any such program or work, and a "work based
on the Program" means either the Program or any derivative work under copyright law:
that is to say, a work containing the Program or a portion of it, either verbatim or with
modifications and/or translated into another language. (Hereinafter, translation is
included without limitation in the term "modification".) Each licensee is addressed as
"you".
Activities other than copying, distribution and modification are not covered by this
License; they are outside its scope. The act of running the Program is not restricted, and
the output from the Program is covered only if its contents constitute a work based on
the Program (independent of having been made by running the Program). Whether that is
true depends on what the Program does.
1. You may copy and distribute verbatim copies of the Program's source code as you
receive it, in any medium, provided that you conspicuously and appropriately publish on
each copy an appropriate copyright notice and disclaimer of warranty; keep intact all the
notices that refer to this License and to the absence of any warranty; and give any other
recipients of the Program a copy of this License along with the Program.
You may charge a fee for the physical act of transferring a copy, and you may at your
option offer warranty protection in exchange for a fee.
2. You may modify your copy or copies of the Program or any portion of it, thus forming
a work based on the Program, and copy and distribute such modifications or work under
the terms of Section 1 above, provided that you also meet all of these conditions:
• a) You must cause the modified files to carry prominent notices stating that you
changed the files and the date of any change.
• b) You must cause any work that you distribute or publish, that in whole or in
part contains or is derived from the Program or any part thereof, to be licensed
as a whole at no charge to all third parties under the terms of this License.
• c) If the modified program normally reads commands interactively when run,
you must cause it, when started running for such interactive use in the most
ordinary way, to print or display an announcement including an appropriate
copyright notice and a notice that there is no warranty (or else, saying that you
provide a warranty) and that users may redistribute the program under these
conditions, and telling the user how to view a copy of this License. (Exception:
if the Program itself is interactive but does not normally print such an
announcement, your work based on the Program is not required to print an
announcement.)
These requirements apply to the modified work as a whole. If identifiable sections of that
work are not derived from the Program, and can be reasonably considered independent
and separate works in themselves, then this License, and its terms, do not apply to those
sections when you distribute them as separate works. But when you distribute the same
sections as part of a whole which is a work based on the Program, the distribution of the
whole must be on the terms of this License, whose permissions for other licensees extend
to the entire whole, and thus to each and every part regardless of who wrote it.
Thus, it is not the intent of this section to claim rights or contest your rights to work
written entirely by you; rather, the intent is to exercise the right to control the
distribution of derivative or collective works based on the Program.
In addition, mere aggregation of another work not based on the Program with the
Program (or with a work based on the Program) on a volume of a storage or distribution
medium does not bring the other work under the scope of this License.
3. You may copy and distribute the Program (or a work based on it, under Section 2) in
object code or executable form under the terms of Sections 1 and 2 above provided that
you also do one of the following:
• a) Accompany it with the complete corresponding machine−readable source
code, which must be distributed under the terms of Sections 1 and 2 above on a
medium customarily used for software interchange; or,
• b) Accompany it with a written offer, valid for at least three years, to give any
third party, for a charge no more than your cost of physically performing source
distribution, a complete machine−readable copy of the corresponding source
code, to be distributed under the terms of Sections 1 and 2 above on a medium
customarily used for software interchange; or,
• c) Accompany it with the information you received as to the offer to distribute
corresponding source code. (This alternative is allowed only for noncommercial
distribution and only if you received the program in object code or executable
form with such an offer, in accord with Subsection b above.)
The source code for a work means the preferred form of the work for making
modifications to it. For an executable work, complete source code means all the source
code for all modules it contains, plus any associated interface definition files, plus the
scripts used to control compilation and installation of the executable. However, as a
special exception, the source code distributed need not include anything that is normally
distributed (in either source or binary form) with the major components (compiler,
kernel, and so on) of the operating system on which the executable runs, unless that
component itself accompanies the executable.
If distribution of executable or object code is made by offering access to copy from a
designated place, then offering equivalent access to copy the source code from the same
place counts as distribution of the source code, even though third parties are not
compelled to copy the source along with the object code.
4. You may not copy, modify, sublicense, or distribute the Program except as expressly
provided under this License. Any attempt otherwise to copy, modify, sublicense or
distribute the Program is void, and will automatically terminate your rights under this
License. However, parties who have received copies, or rights, from you under this
License will not have their licenses terminated so long as such parties remain in full
compliance.
5. You are not required to accept this License, since you have not signed it. However,
nothing else grants you permission to modify or distribute the Program or its derivative
works. These actions are prohibited by law if you do not accept this License. Therefore,
by modifying or distributing the Program (or any work based on the Program), you
indicate your acceptance of this License to do so, and all its terms and conditions for
copying, distributing or modifying the Program or works based on it.
6. Each time you redistribute the Program (or any work based on the Program), the
recipient automatically receives a license from the original licensor to copy, distribute or
modify the Program subject to these terms and conditions. You may not impose any
further restrictions on the recipients' exercise of the rights granted herein. You are not
responsible for enforcing compliance by third parties to this License.
7. If, as a consequence of a court judgment or allegation of patent infringement or for
any other reason (not limited to patent issues), conditions are imposed on you (whether
by court order, agreement or otherwise) that contradict the conditions of this License,
they do not excuse you from the conditions of this License. If you cannot distribute so as
to satisfy simultaneously your obligations under this License and any other pertinent
obligations, then as a consequence you may not distribute the Program at all. For
example, if a patent license would not permit royalty−free redistribution of the Program
by all those who receive copies directly or indirectly through you, then the only way you
could satisfy both it and this License would be to refrain entirely from distribution of the
Program.
If any portion of this section is held invalid or unenforceable under any particular
circumstance, the balance of the section is intended to apply and the section as a whole is
intended to apply in other circumstances.
It is not the purpose of this section to induce you to infringe any patents or other property
right claims or to contest validity of any such claims; this section has the sole purpose of
protecting the integrity of the free software distribution system, which is implemented by
public license practices. Many people have made generous contributions to the wide
range of software distributed through that system in reliance on consistent application of
that system; it is up to the author/donor to decide if he or she is willing to distribute
software through any other system and a licensee cannot impose that choice.
This section is intended to make thoroughly clear what is believed to be a consequence
of the rest of this License.
8. If the distribution and/or use of the Program is restricted in certain countries either by
patents or by copyrighted interfaces, the original copyright holder who places the
Program under this License may add an explicit geographical distribution limitation
excluding those countries, so that distribution is permitted only in or among countries
not thus excluded. In such case, this License incorporates the limitation as if written in
the body of this License.
9. The Free Software Foundation may publish revised and/or new versions of the
General Public License from time to time. Such new versions will be similar in spirit to
the present version, but may differ in detail to address new problems or concerns.
Each version is given a distinguishing version number. If the Program specifies a version
number of this License which applies to it and "any later version", you have the option of
following the terms and conditions either of that version or of any later version published
by the Free Software Foundation. If the Program does not specify a version number of
this License, you may choose any version ever published by the Free Software
Foundation.
10. If you wish to incorporate parts of the Program into other free programs whose
distribution conditions are different, write to the author to ask for permission. For
software which is copyrighted by the Free Software Foundation, write to the Free
Software Foundation; we sometimes make exceptions for this. Our decision will be
guided by the two goals of preserving the free status of all derivatives of our free
software and of promoting the sharing and reuse of software generally.
NO WARRANTY
11. BECAUSE THE PROGRAM IS LICENSED FREE OF CHARGE, THERE IS NO
WARRANTY FOR THE PROGRAM, TO THE EXTENT PERMITTED BY
APPLICABLE LAW. EXCEPT WHEN OTHERWISE STATED IN WRITING THE
COPYRIGHT HOLDERS AND/OR OTHER PARTIES PROVIDE THE PROGRAM
"AS IS" WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESSED OR
IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. THE
ENTIRE RISK AS TO THE QUALITY AND PERFORMANCE OF THE PROGRAM
IS WITH YOU. SHOULD THE PROGRAM PROVE DEFECTIVE, YOU ASSUME
THE COST OF ALL NECESSARY SERVICING, REPAIR OR CORRECTION.
12. IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO
IN WRITING WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO
MAY MODIFY AND/OR REDISTRIBUTE THE PROGRAM AS PERMITTED
ABOVE, BE LIABLE TO YOU FOR DAMAGES, INCLUDING ANY GENERAL,
SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING OUT OF
THE USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT
LIMITED TO LOSS OF DATA OR DATA BEING RENDERED INACCURATE OR
LOSSES SUSTAINED BY YOU OR THIRD PARTIES OR A FAILURE OF THE
PROGRAM TO OPERATE WITH ANY OTHER PROGRAMS), EVEN IF SUCH
HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE POSSIBILITY OF
SUCH DAMAGES.
END OF TERMS AND CONDITIONS
How to Apply These Terms to Your New Programs
If you develop a new program, and you want it to be of the greatest possible use to the
public, the best way to achieve this is to make it free software which everyone can
redistribute and change under these terms.
To do so, attach the following notices to the program. It is safest to attach them to the
start of each source file to most effectively convey the exclusion of warranty; and each
file should have at least the "copyright" line and a pointer to where the full notice is
found.
one line to give the program's name and an idea of what it does.
Copyright (C) 19yy name of author
This program is free software; you can redistribute it and/or modify it
under the terms of the GNU General Public License as published by the
Free Software Foundation; either version 2 of the License, or (at your
option) any later version.
This program is distributed in the hope that it will be useful, but
WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
General Public License for more details.
You should have received a copy of the GNU General Public License along
with this program; if not, write to the Free Software Foundation, Inc.,
59 Temple Place − Suite 330, Boston, MA 02111−1307, USA.
Also add information on how to contact you by electronic and paper mail.
If the program is interactive, make it output a short notice like this when it starts in an
interactive mode:
Gnomovision version 69, Copyright (C) 19yy name of author Gnomovision
comes with ABSOLUTELY NO WARRANTY; for details type `show w'. This is
free software, and you are welcome to redistribute it under certain
conditions; type `show c' for details.
The hypothetical commands `show w' and `show c' should show the appropriate
parts of the General Public License. Of course, the commands you use may be called
something other than `show w' and `show c'; they could even be mouse−clicks or
menu items−−whatever suits your program.
You should also get your employer (if you work as a programmer) or your school, if any,
to sign a "copyright disclaimer" for the program, if necessary. Here is a sample; alter the
names:
Yoyodyne, Inc., hereby disclaims all copyright
interest in the program `Gnomovision'
(which makes passes at compilers) written
by James Hacker.
signature of Ty Coon, 1 April 1989
Ty Coon, President of Vice
This General Public License does not permit incorporating your program into proprietary
programs. If your program is a subroutine library, you may consider it more useful to
permit linking proprietary applications with the library. If this is what you want to do,
use the GNU Library General Public License instead of this License.
Want to help? − Project suggestions
One of the reasons we've made the source code available was so that people could
modify and improve it, port it to new platforms, create new software using VNC, and so
forth. Most of the contributions so far have been ports to new platforms; see the
contribs page for details.
Here are some things that we think would be worth doing. If you would like to sign up
to work on any of these, or if you have any new suggestions, please let us know. We'll
start with the most important:
Improving WinVNC
If you've used both the X server (Xvnc) and the Windows server (WinVNC) you will
notice that the Windows one is considerably slower. There is a simple reason for this.
We have the source code for X, and we know exactly what it is doing. With Windows,
we can get hints by inserting system hooks to monitor messages, but they're only hints
and not all applications use suitable messages, so we often have to poll areas of the
screen just to see if anything has changed. The two main alternative approaches would
be:
writing a pseudo−video driver, which could notify WinVNC with more precise
details
writing a wrapper around the GDI DLLs which would hook into all screen
updates at that level
Security
VNC uses a single TCP/IP connection, so a version which runs over Secure Sockets
should be easy to build. Some users have reported that wrapping the connection using
SSH works well and gives you compression as well. See the FAQ for details. But it
would be nice to have it built−in, not least because SSH for Windows is not free.
People have built the viewer to allow access outward through SOCKS firewalls. See the
contribs page.
Compression
The VNC protocol is fairly efficient in the way it transmits areas of the screen, but on
slow networks a generic compression system would be worth incorporating. The
important requirements are:
It should be possible to switch it off, so that it doesn't increase latency on fast
networks. This would probably mean adding an extra message to the VNC
protocol.
The compression code needs to be easily portable to all platforms including
Java.
The code needs to be freely distributable under the GPL.
Single−application servers
A VNC server doesn't have to be a desktop. In many circumstances it can be useful for
a single application to be a VNC server; we've done a CD player, for example, that can
be controlled by anyone in the room. It would be good to have a simple toolkit to do this
− perhaps the Java Swing toolkit could be extended so that the graphic context object
used for drawing operations could be a VNC server as well...
Steve Cheng has written a GGI 'target' which allows programs using libGGI to become
VNC servers − see the contribs page for details.
You can also get the source code for possibly the most basic VNC server we've written −
rfbcounter.
Attaching to a running X server
As mentioned in the FAQ, the X−based VNC server appears as a separate display for
your applications (typically :1). In theory it would be possible to attach VNC code to a
standard X server so that the applications on the main screen were accessible by VNC,
though with many Unix platforms only the binary of the X server is shipped, making this
more tricky. With XFree86 4.0, it should be rather easier. There is some discussion about
this on the mailing list here.
Other platforms
One of our catch−phrases for VNC is "Any−to−Any"; the basic aim being to access any
user interface from any other user interface. There are several platforms it would be good
to support, but that we can't do here. We used to list several Unix platforms here but
since the release of VNC others have ported it to HP−UX, SGI,. AIX, FreeBSD and
SunOS 4 amongst others − see the contribs page for details.
How about a Windows 3.1 client? The current Win32 viewer is multi−threaded, but if
that aspect were removed it shouldn't be too hard to build for 3.1.
If you have either a client or server running on any new platforms, we'd like to hear
about it. Remember that you should be able to run a viewer on any platform which
supports Java...
Record and Playback
Proxies which do things like recording a session for later playback, extra compression,
scaling, access control, etc are not difficult to create, though you need to keep the latency
low. There has been some work on this − try searching the mailing lists for 'proxy' or
'rfbproxy'.
Other network transports
At the moment VNC uses TCP/IP for the connection between viewer and server, but
that's purely for convenience. It could just as well run over Firewire, USB, RS232,
modems and ISDN lines without IP being there at all. A modem link would be very
handy particularly for remote maintenance; you could dial up PCs which did not
necessarily have an IP address.
We don't, at present, support simple dialup without TCP, for example, because it
involves writing a lot of code to talk to the modem, which would take more time than we
have available at present, and would also expand dramatically the otherwise slim viewer
and server. The right way to do this is probably to provide a scripting extension, so that
the dialup code could be external and easily customised.
VNC − How it works
The VNC Protocol
The VNC protocol is a simple protocol for remote access to graphical user
interfaces. It is based on the concept of a remote framebuffer or RFB. In the past we
have tended to refer to the VNC protocol as the RFB protocol, so you may have seen
this term in other publications. The protocol simply allows a server to update the
framebuffer displayed on a viewer. Because it works at the framebuffer level it is
potentially applicable to all operating systems, windowing systems and applications.
This includes X/Unix, Windows 3.1/95/NT and Macintosh, but might also include
PDAs, and indeed any device with some form of communications link. The protocol
will operate over any reliable transport such as TCP/IP.
This is truly a "thin−client" protocol: it has been designed to make very few
requirements of the viewer. In this way, clients can run on the widest range of
hardware, and the task of implementing a client is made as simple as possible.
Rectangular updates
The display side of the protocol is based around a single graphics primitive: "put a
rectangle of pixel data at a given x,y position". This might seem an inefficient way of
drawing arbitrary user interface components. But because we have a variety of different
encoding schemes for the pixel data, we can select the appropriate scheme for each
rectangle we send, and make the most of network bandwidth, client drawing speed and
server processing speed.
The lowest common denominator is the so−called raw encoding, where the rectangle is
simply pixel data sent in left−to−right scanline order. All clients and servers must
support this encoding. However, the encodings actually used on any given VNC
connection can be negotiated according to the abilities of the server, the client, and the
connection between the two.
The copy rectangle encoding, for example, is very simple and efficient and can be used
when the client already has the same pixel data elsewhere in its framebuffer. The server
simply sends an X,Y coordinate giving the position from which the client can copy the
rectangle of pixel data. This means that operations such as dragging or scrolling a
window, which involve substantial changes to the screen, may only require a few bytes.
Most clients will support this encoding, since it is generally simple to implement and
saves bandwidth.
A typical workstation desktop has large areas of solid colour and of text. Some of our
most effective encodings take advantage of this by efficiently describing rectangles
consisting of one majority (background) colour and 'sub−rectangles' of different
colours. There are numerous other possible schemes. We might use a JPEG encoding
for still images or MPEG for efficient transmission of moving images. An encoding
which uses some kind of caching of pixel data would be good for rendering text, where
the same character is drawn in the same font multiple times. Subsequent occurrences of
the same character would be encoded simply by reference to the first occurrence.
Adaptive update protocol
A sequence of these rectangles makes a framebuffer update (or simply update). An
update represents a change from one valid framebuffer state to another, so in some
ways is similar to a frame of video, but it is usually only a small area of the framebuffer
that will be affected by a given update. Each rectangle may be encoded using a different
scheme. The server can therefore choose the best encoding for the particular screen
content being transmitted and the network bandwidth available.
The update protocol is demand−driven by the client. That is, an update is only sent by
the server in response to an explicit request from the client. This gives the protocol an
adaptive quality. The slower the client and the network are, the lower the rate of
updates becomes. Each update incorporates all the changes to the 'screen' since the last
client request. With a slow client and/or network, transient states of the framebuffer are
ignored, resulting in reduced network traffic and less drawing for the client. This also
improves the apparent response speed.
Input protocol
The input side of the protocol is based on a standard workstation model of a keyboard
and multi−button pointing device. Input events are sent to the server by the client
whenever the user presses a key or pointer button, or whenever the pointing device is
moved. These input events can also be synthesised from other non−standard I/O
devices. On our Videotile, for example, a pen−based handwriting recognition engine
generates keyboard events.
Connection Setup and Shutdown
When the connection between a client and a server is first established, the server begins
by requesting authentication from the client using a challenge−response scheme, which
typically results in the user being prompted for a password at the client end. The
server and client then exchange messages to negotiate desktop size, pixel format, and
the encoding schemes to be used. The client then requests an update for the entire
screen, and the session begins. Because of the stateless nature of the client, either side
can close the connection at any time without adverse consequences.
VNC Clients
Writing an VNC viewer is a simple task, as it should be for any thin−client system. It
requires only a reliable transport (usually TCP/IP), and a way of displaying pixels
(either directly writing to the framebuffer, or going through a windowing system).
We have clients for all the networked display devices we have available at our lab. This
includes the Videotile (the original RFB client), an X−based client (which runs on
Solaris, Linux and Digital Unix workstations), a Win32 client which runs on Windows
NT and 95, a Macintosh client, and a Java client which runs on any Java−capable
browser (including Sun's JavaStation). Members of our lab use these clients on a daily
basis to access their personal computing environments.
VNC Servers
Writing an VNC server is slightly harder than writing a client for a number of reasons.
The protocol is designed to make the client as simple as possible, so it is usually up to
the server to perform any necessary translations. For example, the server must provide
pixel data in the format the client wants. We have servers for our two main platforms, X
(i.e. Unix) and Windows NT/95.
A Unix machine can run a number of Xvnc servers for different users, each of which
represents a distinct VNC desktop. Each VNC desktop is like a virtual X display, with a
root window on which several X applications can be displayed.
The Windows server (WinVNC) is a little more difficult to create, because there are
fewer places to insert hooks into the system to monitor display updates, and a less
clearly−defined model of multiuser operation. Our current server simply mirrors the
real display to a remote client, which means that the server is not 'multiuser'. It does,
however, provide the primary user of a PC with remote access to their desktop.
We have also created simple servers which
produce displays other than desktops, using a
simple toolkit. A "VNC CD player", for
example, generates a CD player user interface
using the protocol directly without any reference to a windows system or framebuffer.
Such servers can run on very simple hardware, and can be accessed from any of the
standard viewers.
VNC − the internal AT&T version
Some of the functionality of the distributed VNC system is limited when compared to
the version we use within AT&T Laboratories Cambridge. This is not because we
wish to deprive the rest of the world of a more sophisticated system, but because we
want VNC to be easy to download and set up, and this extra functionality would make
that process more complicated without offering major gains to most users.
So this information is given partly for interest's sake, and partly to explain some things
you may find in the source code. If there is overwhelming demand for these features we
may make them available, but this is not currently planned.
CORBA
Many of our systems are now linked together using CORBA. We didn't find a
commercial implementation of CORBA which met our requirements, and so we wrote
our own, omniORB, which we have also made freely available. The VNC servers have
a CORBA interface which allows them to be controlled by remote systems. In
particular, the interface exposes the name, address and other details of the server, allows
control of the client connections, can perform remote shutdown, and initiate outgoing
connections.
Outgoing connections
The simplest mode of operation for VNC is when a viewer initiates a connection to the
server, but it can also operate in reverse. Some of our clients can be put into a
'listening' mode where they wait for connections from the server which are initiated and
closed down using the CORBA interface.
External control
This means that desktops can be displayed and hidden under the control of other
systems. In particular we use information from our Active Badge system to display a
users' desktop on a nearby monitor when they click one of the buttons on the badge, and
with our Ultrasound Location System we can make your desktop appear simply by
walking up to a display!
How to keep in touch with VNC developments
The main ways
Just follow these links:
Join a VNC mailing list
Read the mailing list archives
Search the mailing list archives
Read the FAQ
Some other ways
Check these web pages regularly. If you're looking at a mirror or a
downloaded copy of the main VNC site, click here for the master version.
There's a VNC channel in IRC. For information on how to join it, see
this web page.
If you really need to contact the developers directly, send email to
vnc@uk.research.att.com
NOTE: Because of the level of interest in VNC we cannot, in general, reply
to messages, though we do promise to read them all! So if you send a
message to this address, don't necessarily expect anything back. The mailing
list is a much better method of contact; others may have solved your
problem before, and may be able to help if we're not available.
VNCviewer for Java
Because Java applets can only make connections back to the machine from which they
were served, each of the VNC servers actually incorporates a small web server. This
runs on port 58xx, where xx is the display number, and will only serve the Java applet
classes and an HTML page which contains them.
This means that you should be able to point any Java−capable browser at, for example:
http://snoopy:5802/
and you should, after a short pause, be able to connect to your VNC session.
If you are using the X−based VNC server, you may need to specify the directory which
contains the class files in the vncserver script. The Win32 server has the classes
embedded in the server itself.
Java implementations seem to vary widely both in how fast they can read from the
network and how fast they can draw to the screen. It's worth using the Options dialog
to experiment with different encoding schemes for any given network and browser.
Running as an application
You can run the viewer outside a browser using, for example:
java vncviewer HOST snoopy PORT 5902
Note that you need to specify the actual VNC port number and not the display number
or the HTTP port number here! The precise command line will depend on your
particular Java installation.
The VNC mailing lists
If you don't know what a mailing list is or how it works, have a look at this
introduction.
Please, before sending any questions to the mailing list, make sure you've
read the FAQ ! And the documentation for the software you're using.
And read them again. And checked the contribs page and searched the
mailing list archives. There are hundreds of helpful and friendly people on
the mailing list, but they're more likely to remain helpful and friendly if you
don't expect them to do your work for you!
There are 3 different mailing lists − you can choose the one(s) most appropriate for
your level of interest!
vnc−list
Main list for questions, suggestions, bug
reports etc
vnc−list−digest This carries the same messages, but will just
send you one message a day containing all of
the last day's messages. You may prefer this
to the standard list, particularly if your mail
reader won't filter the traffic into a separate
folder for you.
vnc−announce This is for announcements of new versions of
VNC, successful ports to other platforms,
availability of related software. This list is
'moderated', which means that messages sent
here will only appear if we approve them,
which we will do provided they are appropriate
for the list.
Subscribing unsubscribing
You can add yourself to or remove yourself from the mailing lists by sending
commands in the body of an email to majordomo@uk.research.att.com For example:
subscribe vnc−list
This asks for your name to be added to the list. Replace 'vnc−list' with the list of your
choice. More details will then be sent to you by email.
You can remove yourself from the list in the same way, by using 'unsubscribe' instead
of 'subscribe'. You can also just send the command 'help', to get more detailed
instructions. Note that these should be sent to 'majordomo' and not to the list itself! If
your mailer modifies your email address on the way out, you will be subscribed under
this modified name, which can make unsubscribing difficult, particularly if the mailer
changes it to something different at a later time, so if you like you can specify the
email address to be used:
subscribe vnc−list myaddress@company.com
and the same applies to unsubscribing.
Archives
The archives of the mailing lists are available here, and you can search them here.
Go back to 'Keeping in touch'
What's new in the Macintosh VNC package?
See the main VNC History page for an explanation of the version numbers.
Want to get the latest version? Go to the Download Page.
New in Macintosh package 3.3.3beta2
Please see the documentation included in the packages for the latest details.
Changes in Server
Internal changes
Most of the changes won't be immediately noticeable as they are performance/
stability improvements. This version should be faster and better able to deal with
low memory situations than its predecessor. It should also run better under
MacOS 8.6
VNC Controls
The control panel has had a change of appearance and now displays useful
information about the machine and the current connections.
Apple Script Installer
The extension and the control panel can now be easily installed using an installer
script.
Changes in Viewer
Shortcuts
Save the details of your connections as text files with the "Save description.."
option on the File menu.
Logs and Statistics
See what your viewer is doing by choosing "Show Log" or "Show Stats", both
on the File menu.
Custom Keymaps
Remap any key to any other key or sequence of keys, on a global or per
connection basis, using the keymap dialog. Command+G for Global,
command+K for single connection map. Also available from the File and
Window menus respectively. Quickly set a default mouse mapping using "Set
default preferences" from the File menu.
Control, Alt, Delete
Quickly send Ctrl−Alt−Del using the command on the Window menu.
Almost Fullscreen
Fill your screen with a connection by pressing command+F or using the Window
menu.
Connection Options
Change connection options while a connection is active by pressing
command+P. Use this feature to reduce bit depth or change encodings if your
connection is going too slowly.
Scale
View large displays on a small screen by choosing your zoom factor from the
window menu. Also available on command+6,7,8,9,0
Window Arrangement
Easily cascade or tile your windows with command+B and command+T.
Balloon Help
Find out what most buttons do by switching on balloon help. This can be done
from the Help menu or by pressing any of the buttons.
Release date: 11 Jan 2000
New in Macintosh package 3.3.2beta2
Please see the documentation included in the packages for the latest details.
Changes in Server since Alpha 1.1
All preferences can now be set via a control panel − Web interface vastly improved.
Settings and logs should now work.
Changes in Viewer since Beta 1
Better (?) mouse button emulation mechanism (see included docs) Also, both programs
have undergone quite a lot of internal streamlining.
Source code for both programs is now available. Comments, opinions, bugfixes and
requests are welcomed.
Release date: 11 June 1999
New in Macintosh package 3.3.2alpha1.1
• See the documentation for further details.
The server can now use the much faster hextile encoding as well as the 'raw' encoding.
This may introduce some reliability problems − if so you can disable it using the Server
Status Tool.
The server does not now run as a 'background only' application. It no longer generates
its own menu and operates more like a conventional Mac application.
Release date: 2 October 1998
MacVNC − VNC server for Macintosh
Introduction
Use of the Mac server should be straightforward if you are familiar with the concepts
behind VNC, and have a suitable VNC viewer running on another machine. Read
'What is VNC?' and 'Getting Started' if you haven't already done so − they'll give you a
good overview, despite being Unix and Windows−orientated.
Note. This is a beta release. As with all VNC software, you use at your own risk! We
welcome bug fixes and comments.
Requirements:
To use the Macintosh VNC server, you will need:
• A PowerPC−based machine
• MacOS 7.1 or greater. System 7.5.3 is available as a free download from Apple.
• Open Transport (1.1.1 or later). OT 1.1.2 can be downloaded from Apple. If
you are using MacOS later than 8.0 you can use 1.3
Quick Start
• Download the package from the Download page and expand it using a recent
Stuffit Expander.
• Double Click VNC Server Installer
• Restart Machine
• Open VNC Controls control panel (or select Options from the server's File
menu)
• Set password for display 0
• Press Okay, then double click VNCServer
You should now be able to go to another machine which has a TCP/IP connection to
this one, and connect using a VNC viewer.
Documentation
Full documentation is now included with the package. Please refer to this for further
information.
Any comments regarding this program should be directed to the VNC mailing list.
VNC Viewer for Macintosh
Introduction:
Use of the Mac viewer should be straightforward if you are familiar with the concepts
behind VNC, and have a suitable VNC server running on another machine. Read
'What is VNC?' and 'Getting Started' on the VNC web site if you haven't already done
so − they'll give you a good overview, despite being Unix and Windows−orientated.
Documentation
Full documentation is now included with the package. Please refer to this for further
details.
Requirements
To use the Macintosh VNC viewer, you will need:
• MacOS 7.1 or greater. System 7.5.3 is available as a free download from Apple.
• Open Transport (1.1.1 or later). OT 1.1.2 can be downloaded from Apple. If
you are using MacOS later than 8.0 you can use 1.3
Any comments regarding this program should be directed to the VNC mailing lists.
What is a mailing list and how does it work?
A mailing list is a list of email addresses of people who want to correspond about a
particular subject. If you send any message to the list, it gets sent to everybody on the
list, and if you reply to any message, that goes to the whole list as well (unless you
deliberately change the reply address). The VNC list is controlled by a software
package called 'Majordomo' running at AT&T Laboratories Cambridge, and you can
send various commands to majordomo@uk.research.att.com to add yourself to, or
remove yourself from, the list. So the mailing list is a bit like a news server, but the
messages get sent to you rather than you having to go and get them. Until, of course,
you say you're no longer interested.
At the time of writing, about 1500 people are subscribed to the VNC mailing list.
They have joined up because they want to keep in touch with VNC developments.
So it's a good place to ask questions like 'Has anybody got it working on this
platform?' or to say 'I've found a really cool way of using VNC', but it's not a good
place to ask about general problems with your PC networking! You should find a
local expert or some other forum to ask about that.
Back to the 'Keeping in touch' page...
VNC on other platforms & environments
VNC is available for many platforms beyond just the ones we use at our lab. On
this page you will find information about a huge number of other operating
systems and environments for which at least some part of VNC has been ported.
Sometimes there are links to other web sites, often with binaries available. In other
cases you'll find details about how to modify the standard source code to get it to
compile.
We prefer not to distribute binaries directly from our web site for platforms we
cannot test here, but we are happy to put links to sites which do, and we are
grateful to all the contributors. If you port VNC to a new environment we would
be grateful to hear about it; please post such announcements to the mailing list.
Linux RPMs & Debian packages
AIX
Acorn RISC OS
Amiga
BeOS
BSDI
Cygwin32
DOS
EPOC (Psion)
FreeBSD
Geos (eg. Nokia 9000)
GGI
HP/UX
HP Jornada
Java VNC server
KDE
LinuxPPC
MacOS X VNC server
MacOS X Xvnc
MacOS VNCThing (alternative viewer)
MacOS alternative VNC server
NetBSD
NetWinder
Newton
Oberon
OpenBSD
OpenStep/Mach
OS/2
OSF
Palm devices (PalmPilot, Palm III, etc)
RedHat Linux 6.x
SCO OpenServer
SGI Irix
SPARC Linux
SunOS 4.1.3
SVGALIB (Linux without an X server) & Single−floppy Linux
VMS
Windows CE
Windows NT/Alpha
Linux RPMs & Debian packages
Other people have kindly packaged VNC up in RPM and Debian package form; see for
example under 'V' at the RPM archive http://rufus.w3.org/linux/RPM/ and at
www.debian.org . Note that these may not be identical to the standard AT&T distributions,
and may not be updated as frequently, so you should check the main VNC site frequently.
AIX
We've been told you can get VNC for AIX from http://www.bull.de/pub/aix432_frame.html.
Also IBM is including AIX 4.3.3 and AIX5L VNC binaries and source in it's AIX Toolbox for
Linux applications at http://www−1.ibm.com/servers/aix/products/aixos/linux/index.html
Acorn RISC OS
We know of two viewers for RISC OS.
The first, created by simon@bigblue.demon.co.uk, is available from
http://www.bigblue.demon.co.uk/VNC.html .
The second, by Leo White <leo@brighteyes.u−net.com> is at
http://www.brighteyes.u−net.com/ .
A server for RISC OS is also available from http://www.corkes.net/vnc/
Amiga
VVA − a VNC viewer for the Amiga. which was started by Jörg Dietrich is now being
maintained by Denis Spach . See his web page for more details.
Stéphane Guillard − stephane.guillard@steria.fr writes:
I have set up a (incomplete, but usable) Amiga VNC server, which can be found on Aminet
(http://de.aminet.net), as comm/tcp/AmiVNC.lha (ie. here).
This is a work in progress.
BeOS
Andreas F. Bobak writes:
Yesterday, I made a first port of the VNC viewer to the BeOS. I mixed together the Win32
viewer and the Unix viewer and added a bunch of BeOS specific stuff. It basically works but
perfomance does scream. RRE encoding is broken and Hextile encoding has a strange bug, but
it's usable with just CoRRE and CopyRect.
A snapshot of the sources and a x86 binary can be found under http://abstrakt.ch/be/
Cheers
−boby
−− Andreas F. Bobak bobak@relog.ch
BSDI
Kurt Seel <kseel@utcorp.com> writes:
vnc compiles cleanly on bsdi 3.0 (no patches) with the following
pecuiarity − socket.c and httpd.c had to be ifdef'ed like so :
#ifdef __bsdi__
#undef _ANSI_SOURCE
#endif
#include <sys/time.h>
#ifdef __bsdi__
#define _ANSI_SOURCE 1
#endif
It seems to work fine. The switch to xfree 3.3.2 really did some good here!
Cygwin32
Valery Tulnikov has built the server and viewer under Cygwin−32, based on the 3.3.1 patches
by Sergey Okhapkin. This allows you to run the X viewer and server under Win32. Yes, there
are some good reasons why you might want to do this! See http://www.dol.ru/users/valtul/ for
more info.
DOS
Marinos J. Yannikos <mjy@pobox.com> has written a VNC viewer which runs under DOS,
using packet drivers and the Waterloo TCP/IP library along with the Allegro graphics
library. The whole system including the IP stack fits comfortably on a floppy disk. You can
get it from here. .
EPOC (Psion)
We know of two versions of the java viewer to download for EPOC.
Steve Allam < steve.allam@imhotek.com> writes:
Back in February, I first tried the VNC Java client on my Psion Netbook. It worked quite well,
although you had to do nasty command line stuff,− on a psion that doesn't have a command
line by default.
So, I have added two fields; one for the host and the other for the port. I've also added a help
screen that tells you how to move the screen around − i.e the psion screen behaves like a
window on the larger VNC screen.
The whole thing is packaged as an SIS file, and can be downloaded from:
http://www.imhotek.com in the downloads section, at the bottom.
The reason I have taken so long it that I've only just got around to rebuilding the SIS file using
a proper psion UID number! I will be changing the host/port display to make this a combo box,
that saves previous servers, at some point.
Giuseppe Lo Presti < glopresti@inwind.it> has an alternative:
Starting from the Java 1.1 code for the VNC viewer, I packaged an EPOC application so that
the viewer can be run easily from any palmtop which support a Java VM (Psion 5mx &
compatibles). It was tested on the Psion Series 5mx, Series 7, Netbook, and the Ericsson
MC218, and it works with any Internet connection. More information and a downloadable
vncviewer.sis file at http://www.geocities.com/giulopresti/handhelds/vnc−epoc.html
FreeBSD
Joe Evans <evans@ittc.ukans.edu> reports:
I compiled the VNC stuff on FreeBSD 2.2.5, and it seems to work fine. The only compile
glitch was that you need to remove the gnumalloc library from the extra libraries list in order to
do the link step.
Bruce Mah < bmah@ca.sandia.gov > adds:
vnc is now a part of the FreeBSD ports collection...on FreeBSD 2.2.7−RELEASE or newer
with an installed ports collection, the installation process is simply:
cd /usr/ports/net/vnc
make install
Geos (eg. Nokia 9000)
Marcus Groeber < MGroeber@t−online.de > writes:
I have just uploaded the first public version of the VNC Viewer for Geos
to my home page. Have a look at
http://www.mgroeber.de/nokia.htm
It allows you to access VNC servers from machines running Geos 3.0,
which would be either a Nokia 9000/9000i/9110 Communicator, a desktop PC
running NewDeal Office 3.0, or a Brother GeoBook notebook.
Enjoy!
GGI
Steve Cheng steve@ggi−project.org writes:
libGGI is a portable graphics library with a flexible design. See
<http://www.ggi−project.org> for details. The graphics application can betransparently
"retargeted" to different types of displays including X11, Linux svgalib and fbcon. (Win32
soon to come.) The VNC targets adds the VNC protocol to this list. You can even run doom
over VNC! :−)
Standand VNC clients can be used with the VNC target/GGI application as the server.
You can get it here: <http://shell.ipoline.com/~elmert/vnc.tar.gz>
Untar it under degas/lib/libggi of the GGI devel tree. (The stable GGI tree won't work because
of namespace changes,etc.)
A long while ago someone wanted a "stripped−down version of the server part − one that skips
listening/authentication phase and just uses stdin/stdout for the communication (run−once
application)." This is not currently done yet, but I have made it easy to do so. The only
problem is the libGGI application trying to use stdin/stdout.
[The VNC team] have asked me not to distribute this as part of libGGI (yet). (It has GPL'd
code; the other parts are LGPL as the rest of LibGGI). So it won't be
in the GGI CVS tree now.
Bug reports, fixes, and feedback welcome.
HP/UX
Several contributors here. We'll thin these out soon.
Stephen Walton <stephen.walton@csun.edu> who previously posted a detailed message
entitled Building VNC 3.3.3 on HP−UX 10.20 has now also provided some HP/UX binaries.
These are available at: http://www.csun.edu/~swalton/VNC. Stephen sent us this README
Karl Hakimian <hakimian@aha.com> wrote about his experiences building 3.3.2:
HPUX did not go as smoothly as some of the other OS's that have been reported, but I did
manage to get things to compile under hpux 10.20.
First I had to change Xvnc/config/cf/hp.cf
Same as for 3.3.1, I made sure the following were set
#define ExtensionOSDefines −DHPINPUT # −DXTESTEXT1
#define XhpServer NO
#define BuildXInputExt NO
#define BuildPex NO
#define BuildPexExt NO
#define XvncServer YES
I also had to change the following to NO
#define NeedBerklib NO
That got things most of the way compiled. I then could not link Xvnc becuase of several
missing objects
limitNoFile
limitDataSpace
limitStackSpace
Turns out I just needed to include sys/resource.h to two files in the programs/Xserver/os
directory, the following patch takes care of that. (Karl's short patch is here)
Karl Hakimian
hakimian@aha.com
And Mike Cooke writes:
Just to inform you I've managed to build the vnc suite on the following HP box HP−UX
<name> B.10.20 E 9000/879
I had major problems trying to build it with the standard HP tools and after much head banging
and source editing, I decided to forget it and switch to gcc which worked after about the 3'rd
build. I applied the patches as advised in the contrib section on your site.
The only further problem I had was getting the Xvnc server to recognize the correct fonts −
somehow the aliasing config here seems a bit odd, so to get around that problem I just ran the x
font server and it solved all the problems.
And Ulrik Dickow <ukd@kampsax.dk> adds:
This evening I succesfully built vnc−3.3.2r3 on HP/UX 10.20 9000/712 with the
HP ANSI C compiler. I first naïvely tried
xmkmf
make World
cd Xvnc
make World
without any modifications to the source files, but this gave lots of errors
in the Xvnc compilation. Then I changed hp.cf with Karls modifications from
vnc/platforms.html and changed the two C files that his "short patch"
mentions (but in a slightly improved way). This compiled succesfully, with
only harmless warnings. To make it easier for others to build on HP/UX 10,
I've attached the resulting "jumbo patch" to this mail.
It would be even easier if the changes were permanently done to the master
source. They don't affect other platforms than HP/UX, and I guess they'll
help HP/UX 9 and 11 users too, won't they?
To complete the help for fellow HP/UX'ers, the second attached patch modifies
a few lines of the vncserver Perl script to
a) start a CDE session if possible, otherwise fall back to the old code
b) set the `−fp' argument reasonably for HP/UX (the CDE script will add
more to this path).
Due to b), the second patch should *not* be applied to the general
distribution, although a) might be.
If somebody have problems with these plaintext unified diffs, I can provide
context diffs instead, possibly in base64 or uuencoded form. I start out
with unified diffs, since they are much shorter and much more readable in
this case. Apply them in the top vnc source directory like this:
patch −p1 < vnc−3.3.2r3_unix.ukd−kh−hpux−patch
patch −p1 < vnc−3.3.2r3_unix.ukd−hpux_specific−patch.txt
Long live GNU patch(1) (http://www.gnu.org/).
Thanks to ORL for conceiving a great product, and to Karl for finding out how
to compile it on HP/UX 10. BTW, I compiled on a machine with these bundles
installed, as part of an upgrade from 9.0x:
B3393AA_APZ B.10.20.02 HP−UX Developer's Toolkit for 10.0 Series 700
B3898AA_APZ B.10.20.02 HP C/ANSI C Developer's Bundle for HP−UX 10.20 (S700)
B3910BA_APZ A.01.00 HP aC++ Compiler S700
Ulrik's patches are at
http://www.uk.research.att.com/vnc/contrib/vnc−3.3.2r3_unix.ukd−kh−hpux−patch
and
http://www.uk.research.att.com/vnc/contrib/vnc−3.3.2r3_unix.ukd−hpux_specific−patch.txt
Ulrik also pointed out that some HPUX distributions don't include xmkmf and imake, but that
you can get both source and binaries for them from
http://hpux.csc.liv.ac.uk/hppd/hpux/X11/Core/imake−5.06/.
HP Jornada
See Windows CE below
Java VNC server
Tal Liron <tliron@amherst.edu > has written a VNC *server* (not client) in Java, called
VNCj. It is still in an early stage, but essentially works. Available at
http://www.amherst.edu/~tliron/vncj.
KDE
markus.wuebben@kde.org has built a version of the viewer which fits nicely into the X−based
KDE environment.
See http://freshmeat.net/appindex/1998/08/29/904438704.html for details.
Richard Moore rich@kde.org writes:
I've just released the first version of Keystone, a VNC client for KDE 2. The program has a
homepage at http://apps.cx/keystone It is currently available through the KDE CVS snapshots
and will hopefully be included in the next KDE 2 beta release. It is in the kdenetwork package.
This program supercedes the KDE 1 program kvncviewer which you already have listed on
your site.
LinuxPPC
You can get source and binary RPMs for LinuxPPC from
ftp://ppc.linux.or.jp/pub/JRPM/pre−R5/.
MacOS X VNC server
Dan McGuirk <mcguirk@incompleteness.net> writes:
I've managed to hack together a VNC server for OS X, starting with the Xvnc code. (It's a
native server−−it reflects the OS X Aqua desktop over VNC.) It's still got a few rough spots,
but mostly it works pretty well. I put up a web page for it at:
http://www.osxvnc.com/
MacOS X Xvnc
Jeff Whitaker <jsw@cdc.noaa.gov> writes:
I've created a patch that allows Xvnc to compile and run under MacOS X. Details at
http://www.cdc.noaa.gov/~jsw/macosx_xvnc/.
MacOS VNCThing (alternative viewer)
Dair Grant <dair@webthing.net> has written an alternative Macintosh viewer which supports
the Appearance Manager and Navigation Services, which means it looks more attractive on
recent versions of MacOS. You can find it at http://www.webthing.net/vncthing/
MacOS alternative VNC server
Jonathan Morton < chromi@cyberspace.org> has written an alternative Macintosh server.
You can find it at http://www.chromatix.uklinux.net/vnc/
NetBSD
Era eriksson <era@iki.fi> wrote to tell us that information about a NetBSD port for i386 is
available at ftp://ftp.netbsd.org/pub/NetBSD/packages/pkgsrc/net/vnc/README.html
NetWinder
Ralph Siemsen <ralphs@netwinder.org> writes:
I've "ported" the server and client portions of your VNC package to our NetWinder platform
(an ARM−based linux system). There are only minor changes necessary that stem from the
fact that your package tries to rebuild "imake" itself − but is isn't aware of the arm platform.
I've attached the patch below; it is against vnc−3.3.2r3 for unix.
... We'll be providing binaries from our web site for NetWinder owners..
Ralph's patch is available at http://www.uk.research.att.com/vnc/contrib/netwinder−patch.txt
Newton
Steve Weyer <weyer@kagi.com> has created newtVNC (a Newton VNC viewer and server).
See http://members.home.net/saweyer/newton/vnc.htm or http://www.kagi.com/weyer/
Oberon
John Stout and Andre Fischer wrote to tell us of an Oberon VNC viewer. Andre writes:
We have also written an Oberon VNC client that is now used on most of our laboratory PCs. A
first glance can be taken from http://www.oberon.ethz.ch/native/vnc/ .
The software is at ftp://ftp.inf.ethz.ch/pub/Oberon/System3/Contrib/VNC/
Note: The VNC01.Arc is an (proprietary) Oberon archive to install in Oberon. You will find
installation and usage instructions in http://www.oberon.ethz.ch/install.html.
And John said...
...it should give a reasonable system for low end machines. It needs what's called the Gadgets
system so I don't think it would run off a single floppy, but certainly lower requirements than
W95, MacOS, etc. There is a Native Oberon implementation for Linux so I assume it should
work under that.
The link is at:
http://www.oberon.ethz.ch/native/vnc/
OpenBSD
Jonathan Schneider <jschneider@cix.compulink.co.uk> writes that he has built VNC
successfully under OpenBSD 2.5 on both i386 and sparc by doing little more than creating a
symbolic link from /usr/lib/X11 to /usr/X11R6/lib/X11. Either GNU make or BSD make will
work. He needed a 'make −k' to get around the fact that there is no 'vncviewer.man' provided.
He writes:
I've a slight gripe in that on the IPX at least, "make install" puts loads of stuff into
/usr/lib/X11/config* which already existed in /usr/X11R6/lib/X11/config. I imagine that's
desirable if there isn't already an X server installation but scary if there is (and it had used the
correct directory).
OpenStep/Mach
David Young (dwy@picasso.eng.ace.net) writes:
I've written a client for OPENSTEP/Mach (that spiffy NeXT OS) for VNC.
It currently supports display at 24, 12, and 8 bpp, mouse and mostly−functional keyboard input
(ASCII and control characters work; the mapping isn't yet complete and I'm looking for
suggestions on how to complete it), and some other client−side niceties. Encodings other than
raw are on their way, as is NSPasteboard integration.
I'm looking for users (preferably with VNC and OPENSTEP experience) who can bang on this
client on original NeXT boxes, Intel machines, or SPARCstations running OPENSTEP 4.2, or
just people who might find it useful at this early state.
OS/2
Akira Hatakeyama <akira@sra.co.jp>is working on a native PM viewer. You can get it
from http://www.sra.co.jp/people/akira/os2/vnc−pm/index.html
Bosse Nyström bosse@postman.riken.go.jp has built the X viewer for OS/2 using XFree86. He
writes:
I compiled the unix [3.3.1] sources with the attached diffs and got a working viewer under OS2
with XFree86 (and EMX).
I tested it with servers for OSF and Win32 (rev 16), some problems with National
characters for the win server otherwise it works fine.
−− Bosse
You can get Bosse's version from his FTP server at ftp://bfs.riken.go.jp/pub/vnc/, or from
ftp://unixos2.org/pub/unix/internet/vnc/
OSF
We used to have DEC Alpha OSF machines here, but they've all been switched off now, so
we're unlikely to distribute any binaries for any more recent versions of VNC. Nirav Merchant
nirav@arl.arizona.edu has made binaries for OSF 3.3.3 available. He writes:
I have put the OSF binaries, java classes along with the original copyright at:
ftp://public.arl.arizona.edu/pub/sware/vnc−current_osf.tgz
Should work fine with Digital UNIX 4.0.
Since we will be using OSF, I will make sure that the lastest version of VNC OSF binaries are
available at this URL.
Palm devices (PalmPilot, Palm III, etc)
PalmVNC, a viewer for the Palm family of hand−held computers, is now being developed by
Harakan Software − the homepage is at http://www.btinternet.com/~harakan/PalmVNC/
It was originally developed by Vladimir Minenko minenko@icsi.berkeley.edu. You can still
get his version from http://www.icsi.berkeley.edu/~minenko/PalmVNC
RedHat Linux 6.x
If you have difficulties running vncserver on RH6, check at least the following:
• Install the 100dpi fonts from the RedHat distribution if not already installed.
• Check the font path in the vncserver script.
• Check the permissions on /tmp/.X11−unix as mentioned in the FAQ under 'Why can I
only run Xvnc as root?'.
SCO OpenServer
Ben Maizels <bmaizels@analystic.com> sent the patches he used to compile under SCO
OpenServer 5.
You can find his message at http://www.uk.research.att.com/vnc/contrib/sco−openserver.txt .
SGI Irix 6.2 and upwards
See sgi freeware for pre−configured, pre−compiled and inst−packaged VNC server and
viewer.
SGI Irix 6.2
Wolfram Gloger <wmglo@dent.med.uni−muenchen.de> writes:
The following small patch was all that was necessary for me to successfully compile Xvnc on
SGI Irix 6.2 with the N32 binary format. A binary is available at
ftp://ftp.dent.med.uni−muenchen.de/pub/wmglo/.
Wolfram's patch is available at
http://www.uk.research.att.com/vnc/contrib/vnc−irix6.2−patch.txt .
SGI Irix 6.4 and 6.5
Charlie Bond told us that Xvnc compiles OK as long as you do:
cd ./Xvnc ; make BOOTSTRAPCFLAGS=−DFIXUP_CPP_WHITESPACE World
and Ellis Golub posted some useful notes to the VNC mailing list: IRIX 6.5 Build Notes
SPARC Linux
James Hall <jhall1@isd.net> writes:
Just thought I would let you know: VNC 3.3.2r2 compiles just fine on Linux/SPARC. I am
using it now to run Word and Lotus Notes without
having to go back to my office and use my Windows 95 PC.
To compile, I just ran the 'configure' script, then typed 'make'. I think I got a few warnings
during the compile (mostly unused variables) but VNC runs fine. This was on Red Hat
Linux/SPARC 4.x, so I didn't have glibc. Don't know if that makes a difference.
Shinji Hattori writes:
I could not find a VNC package for Sparc Linux. So I made small patch and compiled on
Redhat 5.2 sparc. Xvnc works fine.
You can download his patch here.
SunOS 4.1.3
We've had reports that this buiilds without any problems if you use gcc.
SVGALIB (Linux without an X server) & Single−floppy Linux
One of the very first VNC viewer ports was done by Ganesh Varadarajan and Sitaram Iyer.
They built a vncviewer which runs directly from a Linux console using the SVGALIB library.
You will need to install svgalib and configure /etc/vga/libvga.config for your graphics card.
This can be a bit fiddly − the vgatest program which comes with the svgalib distribution can
sometimes help, and check the setup using another svgalib program such as zgv if you have
problems with the viewer.
The viewer doesn't do any scaling or scrolling, so the desktop you connect to must be of a
resolution supported by your SVGALIB setup. Try creating servers of different geometries and
connecting to them − if your svgalib setup doesn't work for one resolution it may work for
another. Persevere − this has a lot of potential. You can use it to make a very lightweight
graphics terminal with reasonable performance on low−end hardware.
The current version will not generally be able to connect to Windows servers, because it
requests a palette−based display which the Windows server cannot generate.
The sources for svncviewer are here: svnc−0.1.tgz Remember that you will need both X and
the VNC sources on your machine to build it, though you won't need them to run it!
Single−floppy Linux
Karl Heinz Kremer khk@cyberdude.com has created a single−floppy linux distribution which
includes svncviewer. You can now use an old 486 without even a hard disk as an X display.
See http://www.khk.net for details
VMS
A new version of the VNC Viewer for OpenVMS VAX and Alpha is now available.
The VNC333R1VMS011 package contains V1.1 of the OpenVMS VNC Viewer, which is
based on VNC V3.3.3r1. The port to VMS was performed by Hunter Goatley and David North.
It comes as a VMSINSTAL kit, and provides VNCVIEWER and VNCPASSWD images for
both VAX and Alpha. (Note that a VNC server for VMS is not included, though we hope to
look into the possibility of adding it sometime.)
You can get the VMS VNC Viewer using one of the following URLs:
ftp://ftp.process.com/vms−freeware/fileserv/vnc333r1vms011.zip
ftp://ftp.tmk.com/vms−freeware/fileserv/vnc333r1vms011.zip
http://www.tmk.com/ftp/vms−freeware/fileserv/vnc333r1vms011.zip
http://www.process.com/openvms/
Windows CE
The Windows CE viewer is now released! See the Download Page and the Documentation. We
don't currently have the compiler to build for the ARM platform, so if you have an HP Jornada
or similar you may want to get a binary from http://www.conduits.com/ce.
Windows NT/Alpha
Tim Senecal has compiled version 3.3.3r1 for Windows NT/Alpha and made it available at:
http://www.geocities.com/tim_senecal/
Previously John Ross Hunt wrote:
Binaries and VC++ project files are now available for Alpha NT WinVNC3.3.2. You can
download them from: ftp://ftp.lawbulletin.com/vnc/ You will probably have better luck
downloading with a standard FTP client instead of using a web browser (it's a firewall issue).
We plan to upgrade soon, but until then, the old−fashioned way works best.
The VNC protocol
The VNC system is based on the concept of a remote framebuffer or RFB. We
therefore often refer to the protocol underlying VNC as the RFB protocol.
• The protocol specification (Acrobat format)
• The header file rfbproto.h which is used by the system.(Acrobat format)
Additional encodings
Various people have developed their own encodings for VNC. These new encodings
shouldn't break compatibility with other VNC servers and viewers, since a server can
always ignore a request for an encoding, and a viewer never gets sent an encoding it
doesn't request. However it's important that people don't use the same encoding number
for different purposes. Here is a list of the encoding numbers which we know people are
using. If you come up with a new encoding let us know and we'll add it to the list.
• 0 to 5 are standard RFB 3.3 encodings.
• 6 is used by Dave DeBarr's zlib encoding, and also by Tridia's zlib encoding.
• 7 is Const Kaplinsky's tight encoding.
• 8 is the zlib hextile encoding used by Tridia and VNCThing
• 0xFFFFFF00 to 0xFFFFFFFF are currently being used for various
pseudo−encodings, by Const Kaplinsky and Tridia. These may be still
evolving, but this range is best avoided for any other purpose.
rfbcounter
A really simple VNC server.
We thought we'd release the source code of this
just to show that a VNC server doesn't have to
be big and complicated, and that you can display
things other than desktops!
Rfbcounter creates and updates a simple numeric
counter. It does so by generating raw VNC
protocol to produce the digits, and does not have a framebuffer of its own.
The name comes from the fact that the VNC protocol is sometimes referred to as the
RFB protocol. RFB stands for 'Remote Framebuffer'.
The source code for rfbcounter can be found, tarred and gzipped, here. It is a single C
file with a couple of headers, and should compile easily on Linux, and with very little
effort on Win32 and other flavours of Unix.
Usage:
rfbcounter [−clock] display−number width [fg] [bg]
eg. if you run
rfbcounter 5 300
you can then connect a VNC viewer to display 5 on that machine and you'll get a
counter 500 pixels wide. The fg & bg arguments are single byte values specifying
foreground and background colours in BGR233. By prefixing with a zero, you can use
octal to simplify color selection, for example 0007 0100 is bright red on a dark blue
background. If you specify the −clock option, rfbcounter will display the current time
instead of a free−running counter, and update it once a minute.
Note: rfbcounter does not do any pixel translation. Your viewer *must* cope with
8−bit bgr233 format for this to work. On Unix, use the −bgr233 option to vncviewer.
This is not an example of good VNC code. We ignore lots of messages we shouldn't
really ignore, for example. It only uses the RRE encoding, because that's the easiest
way to draw rectangles.
VNC screenshots
These shots show very simple desktops being accessed from a number of different
platforms. Click on the thumbnails for larger images.
An X desktop being viewed from a native PC viewer.
An X desktop being viewed from Microsoft Internet Explorer on a PC.
A Windows machine being viewed from a native X viewer.
A Windows desktop being used from within Netscape on a Unix
machine.
A Unix desktop being accessed from a Macintosh using Java.
A Unix desktop being accessed from a native Macintosh viewer.
A Windows desktop being accessed from a Macintosh using Java.
Making VNC more secure using SSH
VNC uses a random challenge−response system to provide the basic authentication that
allows you to connect to a VNC server. This is reasonably secure; the password is not
sent over the network. Once you are connected, however, traffic between the viewer and
the server is unencrypted, and could be snooped by someone with access to the
intervening network. We therefore recommend that if security is important to you, you
'tunnel' the VNC protocol through some more secure channel such as SSH.
If you are using Unix, this is pretty easy; SSH clients and servers are freely available for
Unix. Clients are also available for Windows, Macs, and other machines, but if you
want servers on these platforms you may need to go for a commercial version, or to route
your connection via a Unix machine (see later). If you're using Java, check out the
modified Java viewer created by Mindbright Technology. There are links at the bottom
of this page to point you in the right direction for all of these things. The rest of this
document refers to the Unix world, though the techniques will be relevant for other
systems. Frank Stajano has contributed a page which describes how he uses a free
Windows SSH client to connect to a Unix server.
Installation
We won't go into details here about how to install SSH. For my Linux machine I found
two RPMs called ssh−clients and ssh−server. I downloaded the source versions, built,
and installed them, and this did almost everything, including the generation of a key for
my machine.
You can get RPMs at http://rufus.w3.org/linux/RPM/ − for other distributions see the
SSH FAQ.
Basic Use
SSH normally just provides you with a 'Secure SHell' − i.e. a login window to a remote
machine. All traffic is encrypted between the two machines using public key encryption
techniques, making it really very difficult for anyone else to spy on it. Once SSH is
installed, you could connect to a machine called 'snoopy' from elsewhere simply by
running the SSH client:
ssh snoopy
(You may need more options depending on your situation). You would then be prompted
for the password of your account on snoopy and you would be logged in, just like a
telnet session, but safer. However, SSH has some nice extra tricks up its sleeve. You
can also request that it listens on a particular port on your local machine, and forwards
that down the secure connection to a port on a machine at the other end. For example,
ssh −L x:snoopy:y snoopy
means "Start an SSH connection to snoopy, and also listen on port x on my machine, and
forward any connections there to port y on snoopy."
Now, the VNC protocol normally uses port 59xx, where xx is the display number of the
server. So a VNC server on a Windows machine, which normally uses display number
0, will listen on port 5900. Most Unix VNC servers will probably use display
numbers 1,2, etc and so will be listening on ports 5901, 5902 and so forth. If you
forward these ports to a remote machine, you can make the remote VNC server appear to
be a server running on your local machine.
So, imagine you had a VNC server running as display :1 on machine snoopy, and you
wanted a secure connection to it from your local machine. You could start the ssh
session using:
ssh −L 5902:snoopy:5901 snoopy
and any references to display :2 on your local machine would actually connect to display
:1 on snoopy, so instead of running a vncviewer:
vncviewer snoopy:1
you could run:
vncviewer localhost:2
and you get the same effect, but with a secure connection.
A quick note if you're using the Unix VNC viewer to connect via SSH: By default, when
the viewer connects to a server on the local machine, it uses VNC's 'raw' pixel encoding
because this generally gives better performance for local access. If this 'server' is
actually an SSHD redirecting the data to another machine, you probably want to override
this using the −hextile option to the viewer, or you will send a lot more data over the
network than is necessary. (On the latest versions of the viewer, use −encodings
"copyrect hextile").
Compression
SSH has another advantage. It can compress the data as well. This is particularly
useful if the link between you and the server is a slow one, such as a modem, but even on
a faster network it can help make up for the fact that the encryption takes a certain
amount of time and so can slow the link down a little. To add simple compression, use
the −C option (or +C in ssh2). For more control, set it up in your ssh configuration files
(see the man page for details). To see how much your data is being compressed, use the
−v option.
If you are on a slow link please read the note in the previous section about connections
to local hosts.
More advanced use
In fact, the connection can be forwarded to a port on a third machine at the other end − it
doesn't have to be snoopy that's running the VNC server, but remember that connections
between snoopy and third machine will not be encrypted, only the one between you and
snoopy. And you can tell ssh to allow incoming connections from other machines and
forward those over the link as well. This can allow you to do really fun things like:
machine windows1
machine
−−−−−−
linux1
Runs vncviewer
secure local Runs ssh to linux2,
connecting to linux1:1,
network fowards local port
actually sees display of
5901 via link to
windows2
windows2:5900
|
|
insecure public
network
|
|
machine
machine
linux2
−−−−
windows2
Runs sshd
secure remote Runs WinVNC
network server as
display 0
To do this, you need to run the following on machine linux1;
ssh −g −L 5901:windows2:5900 linux2
Lastly, remember that if you want to use the Java VNC viewer, you will need to
forward the 58xx ports as well as the 59xx ports. See the FAQ for info on how these
ports are used.
Useful links
• The SSH FAQ.
• Commercial versions of SSH are available from http://www.ssh.fi/
• mats@mindbright.se has created a version of the Java VNC viewer applet which
can connect to an SSH server. See here for details.
• Miroslav Luptak posted detailed information to the mailing list about how he set
his SSH system up on NT using a version built for the Cygwin package. Luis B.
Almeida then posted modified versions for Win95. See the following links:
♦ Miroslav's original
♦ Luis' first modification
♦ A better version
• Various free SSH implementations can be found at
http://www.net.lut.ac.uk/psst/.
• There is also a free SSH client for Windows available at
http://akson.sgh.waw.pl/~chopin/ssh/index_en.html.
Acknowlegements
Many thanks to Axel Boldt, Miroslav Luptak & Luis B. Almeida for information which
helped in the writing of this page.
SSH−protected VNC: the case of the Windows client
and the Unix server
Many thanks to Frank Stajano for contributing this information. You should also read the
general page on Making VNC more secure using SSH if you haven't already done so.
Given that SSH is such a versatile and flexible program, and given that port forwarding is
sufficiently clever that one easily gets one's brain entangled trying to figure out what goes
where, I thought I'd provide a simple step−by−step recipe for the common case in which
one uses a Windows machine as a VNC client to a Unix desktop. I used the free SSH
client which was provided by Cedomir Igaly at http://www.doc.ic.ac.uk/~ci2/ssh/ .
Unfortunately this site seems to have disappeared, but there is a copy of it at
http://linuxmafia.com/pub/ms−windows/igaly−ssh. We've also been told you can get one
at http://akson.sgh.waw.pl/~chopin/ssh/index_en.html or
http://marvin.criadvantage.com/caspian/Software/SSHD−NT/default.php, but we haven't
tried either of them. An alternative is to use the TTSSH extension to Teraterm Pro.
The scenario is as follows: I am not at my laboratory and I am using my own Windows
laptop, called Alice, connected as a guest machine to the network of the university I am
visiting. My laboratory runs an SSH daemon on Unix machine Bob, which is connected
through the lab's network to Unix machine Charlie, which runs my Unix VNC desktop as,
say, Charlie:4. I wish to run a VNC client on Alice that will connect using SSH to Bob
over the insecure wide area network, and from there without encryption to Charlie over
the secure local area network.
Without any of this encryption business, the VNC client on Alice would connect straight
to port 5904 on Charlie.
With the encryption, it will end up connecting to Alice itself (localhost) on an arbitrary
port, say 5954, which SSH will tunnel through Bob to Charlie's 5904.
The difficult question for the newcomer, when faced with something like the opening
dialog box of Cedomir Igaly's SSH Windows client, is where to put all those machine
names and port numbers: are they local or remote forwards? Which host goes where? Do I
need one forward from Alice to Bob and then another from Bob to Charlie, or do I do
everything in one move?
In the terminology of the SSH man page, what we want to do is called a local forward,
because the fake port is created on the local host. The whole route from Alice to Charlie is
covered in just one forward operation, corresponding to the two command line arguments
−L 5954:Charlie:5904 Bob
The first argument is the local forwarding option which says that the local fake port 5954
should be created and connected to the genuine port Charlie:5904. The second argument is
the main non−option parameter to SSH, which tells it the machine to which it should
connect, i.e. the one that runs the SSH daemon.
As far as the Windows client is concerned, Bob goes in the "host name" entry field. The
other three items (5954, Charlie, 5904) should be inserted in the dialog box that pops up
when you press the "local forwards" button, as "local port", "host" and "remote port"
respectively, making up one line of the "forwarded ports" listbox.
Having done this you should type a profile name in the main window, for example VNC,
add your user ID (the one on Bob) in the appropriate field and hit save.
On hitting ok you will proceed to authentication. In the simple case of password based
authentication, you just type in your Bob password and you get a "host is not known"
window with the public key of computer Bob. At this point most impatient users would
just hit "accept key" or "accept once" but you, being careful, will instead verify that it
corresponds exactly to the one you copied from Bob's /etc/ssh_host_key.pub when you
were still at your lab. This will safeguard you from man in the middle attacks where
another computer pretends to be Bob. Once you are happy that the key being presented to
you is indeed Bob's, you can click on "accept key", which causes the key to be recorded
on Alice's known.hosts file, the exact location of which may be specified from the menu
entry "setup/file locations". (Of course you could manually put the key in that file before
even starting the client the first time, which would save you from performing the tedious
verification; but then you would run the risk of getting the format wrong and so on.)
Having done this once, you will never have to perform that verification again.
At that point you should have a shell connection to Bob. You can now iconify that and
start your VNC client on Alice, pointing it at "localhost:54", and everything should work.
Don't forget to follow the official advice on using VNC's hextile encoding rather than raw:
raw is more efficient for a local (same machine) connection, so some viewers will choose
it by default seeing that you connect to localhost; but you are not really on a local
connection, so you should manually select hextile, which is more efficient for connections
that go across the network.
Frank
Note that if you wish to use the Java viewer, you'll also need to forward port 58xx in the
same way as the 59xx examples above, because the Java applet is downloaded via ports
58xx. For the above example, you would forward port 5854 to Charlie:5804 as well as
5954 to Charlie:5904.
Getting Started with VNC
VNC consists of two types of component. A server, which generates a display, and a
viewer, which actually draws the display on your screen. There are two important
features of VNC:
• The server and the viewer may be on different machines and on different
architectures. We expect the most common use to be the display of a Unix X
desktop on a PC, for example. The protocol which connects the server and
viewer is simple, open, and platform− independent.
• No state is stored at the viewer. Breaking the viewer's connection to the server
and then reconnecting will not result in any loss of data. Because the
connection can be remade from somewhere else, you have easy mobility.
So, to get started with VNC you need to run a server, and then connect to it with a
viewer. Get the packages for the platforms you use from the download page, if you
haven't already, and install them. The current VNC software requires a TCP/IP
connection between the server and the viewer, though there is no reason why the
software couldn't be modified to use, for example, RS232 or Firewire. We have internal
versions that use other network transport layers. But for now you'll need to know the
name or the IP address of the server machine.
Most people will be running either a Unix server or a Windows server, though similar
principles will apply to other platforms.
Running a Unix server
A VNC server appears, to the Unix
applications which appear on it, to be a
standard X display just like the one you
sit in front of; but it doesn't have a
physical screen attached. The
applications don't know this. They just
carry on running whether or not a viewer
is connected, in the same way as they
would regardless of whether you were
actually looking at your physical screen.
You can start a new VNC server by
typing:
vncserver
on a Unix machine. (If you're sitting at
a PC you may need to telnet to the Unix
machine to get a command shell into
which you can type this.) The
Running a Windows server
Install the Windows server, WinVNC, by
running the Setup program included in
the distribution. This will create a VNC
group in your Start Menu.
Install the default registry settings using
the option in the VNC group.
Run the WinVNC server. If this is the
first time you've used WinVNC on this
machine you'll be prompted to set a
password, which you'll need when you
connect to the machine from a remote
vncserver program is a Perl script
which you may need to edit to set up the
directories appropriate for your local
installation.
If you haven't run a VNC server before
you will be prompted for a password,
which you will need to use when
connecting to this server. All your
servers will use the same password, and
you can change it at a later date using
vncpasswd
With a normal X system, the main X
display of a workstation called 'snoopy'
is usually snoopy:0. You can also run as
many VNC servers on a Unix machine
as you like, and they will appear as
snoopy:1, snoopy:2 etc, as if they were
just additional displays. You can cause
applications to use them by setting the
DISPLAY environment variable to the
VNC server you want, or by starting the
application with the −display option.
For example:
xterm −display
snoopy:2 &
Normally vncserver will choose the
first available display number and tell
you what it is, but you can specify a
display number if you always wish to
use the same one:
vncserver :2
Nothing will appear immediately as a
result of starting a server. To see
anything you need to connect a viewer to
the server. See below. The server will
generate a log in your ~/.vnc
directory. If you have problems at this
stage, see the full documentation and the
FAQ .
Killing a Unix server
You can kill a Unix VNC server using,
for example:
vncserver −kill
location. Normally you'll want to leave
the other options on their default
settings. (Note that the default display
number is 0 on a PC. You'll need to
specify this to the viewer when you
connect.) Click OK and the server
should be running. It will install a small
icon on the system tray, and by
right−clicking on this you can control
most aspects of the server. (Full
instructions for installing and running the
WinVNC server can be found under the
main documentation.)
You can now go to another machine and
connect a viewer to the server. (see
below)
:2
Running a viewer
When you run the viewer, you need to specify the name of the server and the number of
the desktop. If, for example, you have started a server as display 2 on a machine called
'snoopy', you can start a viewer for it by typing:
vncviewer snoopy:2
With the Windows viewer, you can run it from the command line, but you will more
typically run it from the VNC group on the Start Menu. In this case, you will be
prompted for the host name and display number:
Enter it and click OK, and you will be prompted for your password, after which you
should see the remote display. (If you are connecting to a Windows or Mac server, the
display number will be 0, unless you have explicitly changed it).
If the machine running the server does not have a proper DNS entry, you probably won't
be able to use the name and will have to replace snoopy:2 with something like
192.168.1.2:2 . You can get round this on most platforms by creating a 'hosts' file
which maps names onto IP addresses. Consult your local guru for help with this.
Using a web browser as a viewer
The VNC servers also contain a small web server. If you connect to this with a web
browser, you can download the Java version of the viewer, and use this to view the
server. You can then see your desktop from any Java−capable browser! (unless you are
using a proxy to connect to the web). The server listens for HTTP connections on port
5800+display number. So to view display 2 on machine 'snoopy', you would point your
web browser at:
http://snoopy:5802/
The applet will prompt you for your password, and should then display the desktop.
That's it! For more details see the documentation. The answers to lots of common
questions can be found in the FAQ .
Improving VNC performance with 'twm'
This may give some ideas for other Window managers as well, though the syntax will be
different. See the FAQ for other suggestions on speeding up a VNC session.
• Don't make the desktop deeper than you need to.
• disable the dithering on the title bar by using a solid bitmap.
In .twmrc put:
Pixmaps
{
TitleHighlight "solid"
}
and put a bitmap file 'solid' on the bitmap path containing:
#define solid_width 4
#define solid_height 1
static char solid_bits[] = {0x0f};
You can set the X bitmap path − I have in my .Xresources:
*bitmapFilePath:/home/qsf/bitmaps
• make the scrollbar solid
*Scrollbar*thumb: solid
*Scrollbar*foreground: grey
• remove the subtle modification of the window border when it gets
focus. In .twmrc:
NoHighlight
• use click−to−focus instead of point−to−focus (does twm allow this?)
Then the window doesn't change when your mouse just happens to move over it.
VNC Videos
We have low−quality versions of the VNC publicity video available for download.
We hope to stream them, eventually, but for now you can get them from the FTP server
and play them from your local disk. You probably want to right−click the links below to
tell your browser to save a local copy.
The video is just over 8 mins long, and is encoded using RealVideo[tm]. You'll need
version 5.0 or later of the RealPlayer to play them; you can get it from www.real.com.
vnc−28k.rm is only 1.6Mbytes and is very low quality as it
was encoded for streaming over a 28k modem. But it
shouldn't take long to download!
vnc−56k.rm is 2.8Mbytes and is
slightly higher quality.
vnc−128k.rm is just over 6M and is
quite reasonable quality.
There's also vnc−300k.rm (19M) for those who are really keen and have a fast link to
our site!
There is no virus in WinVNC.
Some versions of McAfee virus scanning software reports that there's a "new backdoor"
virus in WinVNC. This is due to McAfee's Heuristic Virus detection mistakenly
identifying WinVNC as a virus. McAfee's website has a suitable extra.DAT to fix the
problem until the next full DAT release.
VNC History
What do the version numbers mean?
The overall VNC system has a version number, currently 3.3.3 at the time of writing,
which means it's the third version of the system based on protocol version 3.3.
We add features and fixes to different platform packages at different times, so there may
be several Revision numbers within a given version. Revision 5 of version 3.3.2 is
written 3.3.2R5.
We don't necessarily release all our revisions, especially with the Windows package, so
there are often gaps in the numbers. We tend to number the Windows packages to
match the version of the WinVNC server, so if you want to find out which package
you've got, have a look at the About box of the server.
Platforms
• Change history for the Unix and Java packages
• Change history for the Windows package
• Change history for the Macintosh package
Want to get the latest versions? Go to the Download Page.
VNC people
The following people, and no doubt others we've forgotten, are working on VNC or have
contributed to the development at some point in its history, and we are grateful to them
all. You can click on the pictures to go to their home pages.
Please don't mail these people directly with VNC problems. See the Keep In Touch page
for more contact info.
Tristan Richardson
Quentin Stafford−Fraser
James Weatherall
Ken Wood
Andy Harter
Charlie McLachlan
Paul Webster
What's new in the Windows VNC package?
See the main VNC History page for an explanation of the version numbers.
A more complete list of changes can be found in the files called History.txt included in
the source packages of both the viewer and the server. This is just a summary.
New in Windows package 3.3.3r9
Fixed a bug introduced in R8 which broke the AuthHosts feature.
The WinVNC.log file is now moved to WinVNC.log.bak on startup if already present.
Release date: 19 March 2001
This package consists of version 3.3.3r8 of the server and 3.3.3r3 of the viewer.
WinVNC server
Major
• Clipboard handled correctly.
• AuthHosts filter now behaves correctly.
• Screen−saver killing code is safe => re−introduced.
• RemoveWallpaper feature added.
Minor
• Socket leak in AddClient removed.
• QuerySetting works even with no AuthHosts string set.
• Do you really want to quit? dialog never appears in service mode.
Tweaks
• Query connection dialog is now more conspicuous.
Fixes to VNCHooks
• VNChooks removes custom window properties on exit. Fixes Delphi problems!
• Added HooksType() function to VNChooks.
Release date: 15 March 2001
Fixed winsock−related bug which caused R6 server not to work under unmodified
Windows 95.
Release date: 26 May 2000
The viewer has not changed in this release. The server has been upgraded to 3.3.3r6.
WinVNC server
As always, see the History.txt file in the source release for full details. This revision
incorporates:
• Many bug fixes and some speed improvements (particular thanks to Will Dean)
• Some protection against Denial−of−Service attacks.
• The QuerySetting QueryTimeout registry options which allow the user of a
desktop to be prompted before others can connect.
For details see the WinVNC documentation.
Release date: 18 May 2000
This package consists of version 3.3.3r2 of the server and 3.3.3r2 of the viewer.
Windows VNCviewer
Changed threading to fix problems on some SMP machines.
The −register option now stores quoted pathnames, so the −config option should
work better when filenames have spaces in them.
Config files now store details about the VNC encodings selected.
Fixed bug which could cause a crash in rare circumstances when given an empty server
name.
WinVNC server
Added −connect option which causes a currently running server to initiate an outgoing
connection. (Equivalent of 'Add New Client' menu option.)
Outgoing connections are now in 'shared' mode.
Improved the way preferences are loaded, reducing the number of times clients will be
disconnected during loading.
Added experimental option to disable local keyboard and mouse when remote
connections active. Works on NT only.
Fixed bug in the 3.3.3 server which was stopping the Java viewer from working. This
package consists of version 3.3.3r1 of the server and 3.3.3 of the viewer.
Release date: 15 September 1999
New in Windows package 3.3.3
The documentation has been updated. Please read it again if you're upgrading.
The Windows VNC package is now built using Visual C++ 6.0. We recommend that you
use the same or later if you wish to rebuild it from the source code.
Windows VNCviewer
Changes to thread scheduling and locking which should improve reacitivity of local GUI
(menus, scrollbars etc) on slow links.
Information about the current connection including, optionally, the password, can be
saved to a '.vnc' file, which can then be specified on the command line. You can use the
IniFileMapping facility of NT if you would prefer these to be saved in the registry. Also
an option to register '.vnc' files as belonging to vncviewer, so that you can double−click
on them.
Fairly primitive scaling facility allows you to see desktop at different sizes.
In Connection dialog, display defaults to :0 if not specified, so when connecting to
WinVNC servers you can usually just type the machine name.
Command line options to change the appearance of the local cursor.
3−button emulation (create a middle−button click by pressing left and right together) is
now the default, so an option has been added to disable it if required.
Updated copyright notices and About... boxes.
WinVNC server
An 'Add New Client' option has been added to the server menu, which initiates a
connection to a listening VNC viewer (one started with the −listen option). This allows
you to send your desktop to a helpdesk, for example, without them needing to know your
password or IP address.
The WinVNC settings for the currently logged in user are now correctly loaded even
when running as a service under Windows NT, provided a small 'helper' app is run when
the user logs in. This helper is installed automatically when WinVNC is run as a service.
Limited client IP−address based filtering of connections is now supported, through the
new Machine−Local registry setting "AuthHosts".
Incoming connections may be restricted on a server−wide basis to being accepted only
on the local−loopback interface, through the "LoopbackOnly" registry setting.
Nagle's algorithm is now disabled on all WinVNC connections, decreasing latency
considerably in many cases.
Dynamic priority adjustments are made to WinVNC threads to increase performance
without impacting other running applications.
The machine's name is now included in the title of the web page used to access the Java
version of the VNC client.
Timestamp information is now added in the debug log output.
MAJOR FIX: Two significant handle leaks which could occur when running as a service
under Windows NT have been fixed.
MINOR FIX: The "LockSetting" registry option is now correctly loaded.
MINOR FIX: A bug which would sometimes cause an Application Fault when stopping
a service−mode WinVNC has been fixed.
Release date: 11 September 1999
New in windows package 3.3.2 R6
The R6 package consists of:
• Revision 6 of the WinVNC server
• Revision 11 of the Windows VNC viewer.
Windows VNCviewer (R11)
Added middle−button emulation for two−button mice by clicking both buttons
simultaneously. This need to be turned on from the command line or the Options
dialog. Many thanks to Randy Brown <rgb@inven.com> for providing this very useful
feature, which we've been meaning to do for some time! A few new command line
options to control this.
You can start connections in full−screen mode by specifying −fullscreen on the
command line.
You can disable transfer of the clipboard in both directions from the Options dialog or
the command line.
WinVNC Server (R6)
Fixed a bug in polling of DOS−prompt boxes under Windows 95.
Fixed minor protocol non−compliance bug with regard to clients which fail to send
SetPixelFormat messages.
Release date: 22 January 1999
See the documentation for further details. The R5 package consists of:
More internal restructuring so that more code can be shared with the Windows CE
viewer. Fixed some resource/handle leaks.
Added a 'Request Refresh' option to the menu.
WinVNC Server (R5)
The WinVNC registry settings have now been reorganized and improved to allow better
control over individual user's capabilities when running it, whether as an application or
as a service. Full details in the documentation.
Mouse events are now produced to reflect those sent on the wire, regardless of whether
the mouse buttons would locally be swapped by the OS. This means that left− or
right− handedness is a client−side feature.
If you hold the mouse over the WinVNC icon in the System Tray, it will attempt to
display the local IP addresses of the machine. Useful if these are dynamically allocated.
Some of the more important fixes:
If a client chose to send multiple update requests to the server without waiting for the
update data, then the server would only consider the last request. This behaviour was
incorrect and has now been fixed. As a result, WinVNC will now work properly with the
PalmVNC viewer.
The HTTP server section of WinVNC could end up listening on the wrong port number
when automatic display number selection was in use. This is now fixed.
A small memory leak in the WinNT side of the vncService::CurrentUser function has
been fixed. Although the amount lost on each execution was small, the function is
called often enough for this to cause problems over long periods of use.
Fixed a potential lockup when the Properties dialog is closed by the OS rather than by
the Ok or Cancel buttons.
See the documentation for further details. The R2 package consists of:
Added full−screen mode. Many thanks to Greg Hewgill <greg@hewgill.com> for
providing much of the basis for this.
Creates a palette for much better colour on 256−colour clients.
Added −viewonly and −restricted switches.
In addition to Ctrl−Alt−Del, you can send individual Ctrl−up, Ctrl−down, Alt−up and
Alt−down keystrokes from the menu.
Updates are not requested from the server while iconified.
The initial state of the clipboard is not sent on connection as it was in the past.
Substantial source code restructuring.
The local 'dot cursor' is only used when the viewer window has the focus.
The modifiers are all raised when losing focus.
In 'listen' mode, the tray icon is checked every 15 seconds and restored if not there. If
your Explorer dies as often as mine, you might appreciate this.
Fixed session dialog box so longer host names could be entered.
WinVNC Server (R2)
Palette−based clients (such as the Linux SVGAlib client) are now properly supported by
WinVNC.
Local−loopback connections, from viewers on the same machine as the server, can now
be enabled.
Null−password connections can now be enabled.
WinVNC server logging features are in place, allowing remote connections and error
conditions to be logged.
More of the keypad keys are now recognised from remote clients. This means that
Ctrl−Alt−keypadDel will cause the server to perform the standard Ctrl−Alt−Del
procedure.
Extra hook code is in place to make updates of pop−up menus, etc, work better.
The new ConnectPriority option is available, on a per−machine basis, to indicate how the
server should deal with multiple non−shared connections.
Automatic display number selection is now available.
The CopyRect optimisations are now better handled.
Screen update notifications are now handled more efficiently.
The WinVNC −kill command will now wait until any old copies of WinVNC have quit
before attempting to continue.
Release date: 23 July 1998
New in 3.3.2
The MRU (most recently used) list on the viewer now works much better. Extra
command line options allow for more debugging output.
WinVNC now returns explanatory message text to the viewer when an incoming
connection is refused because of an empty Password field or a local−loopback
connection.
Mouse movements are now pre−processed by the Desktop handler and only passed to the
Client handlers immediately before an update is triggered, resulting in less overhead and
therefore lower latency.
If a client requests the CopyRect encoding in the list of encodings it supports then
WinVNC will now use the CopyRect primitive when doing simple tasks like
dragging windows. This still has some flaws in it but generally improves performance.
Updated the Java classes exported by WinVNC to include the Send−Ctrl−Alt−Del
button.
Added a "−about" option, which will cause an _already running_ copy of WinVNC to
display its About box, making it easier to check that you have the latest version!
Added version information to the VNCHooks Dynamic Link Library. This can be
viewed selecting the properties dialog for the vnchooks.dll file.
New in 3.3.1 Revision 19:
MS−DOS applications can now by typed into! Thanks to Gur Stavi for pointing out that
WinVNC didn't previously generate keyboard scancodes for key events. This means that
the COMMAND.COM and EDIT programs, for example, may now be used through
VNC.
Lookup−table based colour translations are now incorporated into WinVNC. This means
a small increase in memory usage but delivers a noticable performance boost on most
screen formats.
1, 4 and 8 bit palette−based local displays are now handled directly by WinVNC,
resulting in a significant performance boost, although there are currently problems with
the palette layout being confused in places.
More intelligent culling of unchanged regions of the screen from the list of rectangles to
be sent has been implemented. Generally, this isn't noticable but over low−bandwidth
links, it should have a significant effect.
WM_ENDSESSION is handled properly for full system shutdown.
WM_DISPLAYCHANGE is now handled, so that when the display resolution changes,
all remote VNC users are disconnected, to prevent corrupt display updates.
The mouse cursor rendered to VNC clients is now correct most of the time.
When installing WinVNC as a service, quotes are now placed around the executable's
name, to avoid problems if there are whitespace characters in the path. (Common
because of installing to "program files\orl\vnc").
The command−line options available have changed slightly − see the WinVNC
documentation for details.
The Java viewer class files have been updated. The new classes are slightly more
compatible with borderline Java VMs.
Deferred update messages are no longer removed from the application's message queue
behind its back, hopefully resulting in more reliable behaviour & fewer lock−ups.
The screen−saver is now not disabled when WinVNC is running in service mode and a
connection is made, to avoid a potential race condition. This will be handled better in a
future revision.
WIN−NT ONLY:
When running as a system service, WinVNC no longer disconnects all remote VNC
connections whenever the current desktop changes.
Corrected some problems with shift−key release code confusing Windows NT.
WIN−95 ONLY:
WM_USERCHANGED is handled, so that when no user is logged in, the machine−local
password is used, otherwise the user's own VNC password and settings are used. (This
assumes that Windows 95 is set to Multiple User Profile mode)
Under Windows 95, WinVNC running in service mode no longer crashes when told to
stop by the −kill or −remove options.
The main text area of the Windows 95 console will now be polled by WinVNC properly
when the Poll Console Windows Only option is set.
The big change is that WinVNC can now run as a service under both 95 and NT, so you
can log in and log out remotely. This is still under development, so please excuse some
rough edges! For more details see the WinVNC documentation. It will still run in its
conventional mode if preferred.
The viewer now has a 'Send Ctrl−Alt−Del' option on the menu for use with the above.
Unix viewers can generally send this anyway, as can Java viewers on platforms which
don't use it locally.
A bug which often caused the entire screen to be transmitted twice to a connecting client
is now fixed.
The general Shift, Alt and Control−related bugs, affecting non−UK/US keyboard layouts
in particular, are now corrected, with one important exception:
When typing into an application set to use international keyboard layout A, using
WinVNC run with keyboard layout B, problems may be experienced because WinVNC
sets the wrong Shift key states to generate particular characters. As long as only one
keyboard layout is used consistently across all applications, this shouldn't be a problem.
The viewer now generates AltGr international characters according to the keyboard
layout in use. The right Ctrl key can be combined with these so that on a German
keyboard, for example, AltGr−Q gives @ and RightCtrl−AltGr−Q gives Ctrl−@. Dead
keys are not properly supported yet.
The Caps−Lock, Num−Lock and Scroll−Lock keys are now correctly disabled whenever
a new client connects, if that client has keyboard input enabled.
Left and right versions of the Shift, Alt and Control keys can now be distinguished on
NT if the viewer distinguishes.
The right and middle mouse buttons can now be set to trigger window updates, although
by default these options are not used. This can be handy for things which use right−click
for something other than popping up a menu.
WinVNC − The Windows NT VNC server
WinVNC is a VNC server that will allow you to view your
Windows desktop from any VNC viewer. Because
Windows in its present, standard incarnation, only supports
a single graphical user being logged in at any one time,
WinVNC makes the existing desktop of the PC available
remotely, rather than creating a separate desktop as
happens with the Unix server. It is only fair to emphasise this: VNC does not make an
NT machine into a multi−user server in the same way that Citrix−based software, for
example, does. A single NT machine can therefore be accessed by multiple users, but if
they all connect at the same time they will all see the same desktop!
On the other hand, WinVNC will run on Windows 95, Windows98, Windows NT 4.0,
Windows 2000 and on any future Win32−based systems, without the need to replace any
system files or run any OS−specific versions of the program. It is a standard application
that can be run from the Start... menu and closed down just as easily.
WinVNC can also be run as a service, which means that you can log in remotely, do
some work, and log out again. We recommend that you run it in this mode. See below for
more details.
And, of course, WinVNC is free. We hope that making the source code available will
enable others to suggest improvements to any and all aspects of WinVNC.
If you haven't yet read the 'Getting Started' page, you might like to do that first to give
you the general concepts.
Installation
WinVNC is simple to install and to use:
1. Run the WinVNC setup program. (If you received VNC as a ZIP or TAR
archive, you'll need to unpack it first!)
2. Select Install Default Registry Settings from the WinVNC folder in the Start
menu. This will install the default hooks settings, which are tweaked to cope
with some common, uncooperative applications, such as the clock. This is not
vital, but will make things run a bit better. See later for more information about
the registry settings.
3. WinVNC can now be run (in Application mode) from the Start menu.
Alternatively, you can use the Start−>Settings−>Taskbar menu to add a shortcut
to your Startup group, which will cause WinVNC to be run every time you log
in. NOTE: After some initial tests, we suggest you run WinVNC as a service
instead of as an application. See below for more information.
Using WinVNC
On starting, WinVNC will add a small, green version of the VNC icon to the system task
bar. Clicking on this icon with the right mouse button will cause a menu to be displayed,
with the following options on it:
• Properties − This will cause the User Properties dialog to be displayed, allowing
the user to change various per−user WinVNC parameters.
• Add New Client − This allows outgoing connections to be made from the server
to any "listening" viewer. The name of the target viewer machine can be entered
in the dialog. Connections created this way are treated as shared, as of 3.3.3R2.
See also the −connect option below.
• Kill All Clients − This will disconnect all currently connected clients from the
server.
• About WinVNC − This should be obvious!
• Close − Shutdown the server.
Moving the mouse over the icon should cause the IP addresses of the local machine to be
displayed, if they can be discovered at that time.
You can connect to the server from another machine using a VNC viewer, as described
in the Getting Started page.
WinVNC Properties
The following options are available from the Properties dialog.
Incoming Connections
• Accept Socket Connections − The server normally accepts direct, socket−based
connections from the vncviewer program. Clearing this tick−box disables direct
connection to WinVNC, so that only the CORBA interface used by our internal
version may be used to start a connection. (See AT&T internal version info).
For the public version, clearing this will disable any incoming connections.
• Display Number − This allows the user to specify the display number which the
server will use. There is normally no need to change this from the default of
zero.
• Auto − This tick box indicates to WinVNC whether it should use the display
number specified in the Display Number box, or whether it should use the first
display number not already in use on the server machine.
• Password − Incoming connections must be authenticated to verify that the person
connecting is allowed to connect to this machine. This text box allows your
password to be specified for authentication.
• Disable Remote Keyboard & Pointer − Any new incoming connections will be
able to view the screen but not send any input.
• Disable Local Keyboard & Pointer − This is experimental, and works on NT
only. If selected, then the local keyboard and mouse will be disabled during a
connection. Useful if you want to log in to a machine from elsewhere and don't
want passers−by to be able to use your session.
Update Handling
Note that clicking in a window will generally cause it to be updated, so if you have
certain applications which don't update very well, try this! The default update handling
settings should be the right ones for most people, and in general you will slow things
down by changing them, so don't do this unless you have applications which cause
problems.
• Poll Full Screen− Some applications are incompatible with the methods currently
used in WinVNC to trap screen updates. For this reason, it is sometimes useful to
be able to poll the entire screen in order to check for changes, sacrificing
performance for accuracy.
• Poll Foreground Window − Polling only the currently selected window for
changes is less CPU intensive than full−screen polling and often gives similar
results, for example when using the Command Prompt, which is not normally
compatible with WinVNC.
• Poll Window Under Cursor− A variation on Poll Foreground Window, this
option causes the window under the mouse cursor to be polled for changes. Both
options may be enabled simultaneously if required.
• Poll Console Windows Only− When this option is set, the only windows which
will be ever be polled are Command Prompts. This works well in conjunction
with Poll Window Under Cursor, to use polling only when the cursor is over a
console window.
• Poll On Event Received Only− When this option is set, the screen will only be
polled for updates when a mouse or keyboard event is received from the remote
client. This is provided for low bandwidth networks, where it may be useful to
control how often the screen is polled and changes sent.
The user's settings are saved into the user−specific section of the registry when WinVNC
quits, meaning that they will be used next time you run WinVNC.
Running WinVNC as a service
WinVNC can now be made to run as a service process under both Windows NT and
Windows 95/98, by following the instructions outlined below. This allows you to
connect to a machine which has nobody logged in to it, for example. On NT, you can
also send Ctrl−Alt−Del to the server when it's running as a service, allowing you to
unlock a locked workstation, for example. Note that in 'service' mode, many features are
changed on a per−machine rather than per−user basis. You can access the per−machine
'Defaults' from the 'Administrative Tools' section of the VNC Start Menu group. The
following 'features' should also be pointed out:
Windows NT 'features':
• WinVNC will attempt to correctly identify the user and locate their preferred
settings in the registry (unless AllowProperties has been used to disable this −
see below). To do this requires that a helper application be run when a user logs
in, which will pass appropriate information to the main service portion of
WinVNC. If the helper fails to run for some reason then WinVNC will continue
to operate but it won't know who is logged in, so its settings will be based on the
local−machine and default−user settings only.
Windows 95 'features':
• Whether or not the VNC password is set per−machine or per−user depends on
the settings in the Passwords section of the Control Panel. If Win95 is set to use
a different set of registry values for each user then when a user logs in, the
password will change from the per−machine VNC password to that user's VNC
password. If Win95 is set to use the same settings for all users then the
per−machine VNC password will always be used.
General features for both Windows NT and Windows 95:
• Anything which causes the Windows VNC server to change screen resolution
will also cause all viewers to be disconnected, and you'll need to reconnect.
Logging in can sometimes do this if the user has a different screen setup from the
system default.
• If a user has specified a display number which is different from the default used
by the service, viewers will be disconnected as that user logs in.
• When WinVNC is running as a system service, no user−level copies can be run
at the same time.
Here's how to get it running as a service, assuming you've already installed it. Under
Windows NT you need to have administrator privileges on the local machine, so log on
as administrator if your account doesn't have these.
1. Either:
Select "Install WinVNC service" from the WinVNC Administrative Tools
section of the Start menu.
Or:
Open a Command Prompt and run WinVNC with the −install option. eg:
D:\> C:
C:\> cd "\Program Files\ORL\VNC"
C:\Program Files\ORL\VNC> winvnc −install
2. Windows 95 : The WinVNC service is now running and is installed to run
whenever the system boots up into Windows 95.
Windows NT : The WinVNC service is installed and set up to run whenever the
machine is booted into Windows NT but IS NOT YET RUNNING! It will run
when the machine next reboots. If you want to start it immediately, you can use
the Services section of the Windows control panel, or "net start" from the
command prompt:
C:\> net start winvnc
The service should now be running, but won't know that you are logged in until
you log out and back in again. It won't yet appear on the taskbar. You can run the
WinVNC Service Helper from the start menu to tell it who you are!
3. If you wish to change the WinVNC settings (eg. password) when it is not visible
on the taskbar, you can use Show User Settings from the WinVNC section of the
Start menu.
If you wish to change the default settings used by the service when there's
nobody logged in, or no user−specific settings in are use, use Show Default
Settings from the Administrative Tools subsection of the WinVNC Start Menu.
See below for more information about defaults.
NOTE : Windows 95 : If Win95 has been set to use different settings for each
user then the settings used are those of the currently logged in user. If no user is
logged in or Win95 is set to use the same settings for all users then the settings
used are the Default user settings and are stored per−machine, rather than on a
per−user basis as is done when running WinVNC normally. (Under Win95,
pressing Cancel on the login dialog gives access to the Default user settings.)
4. When you wish to install a new version of WinVNC or simply wish to remove
WinVNC from your machine, you must first remove the service from the
system, using the WinVNC Administrative Tools or the −remove command−line
option.
NOTE : A message about failure to remove the service usually indicates that it
was not installed in the first place!
Command−line options
The full command−line options available are as follows. You probably won't need
anything other than those listed above unless you're a real VNC power−user!
−run
Causes WinVNC to run normally & ignore rest of command−line.
−install
Installs the WinVNC service and continues reading the command−line.
−remove
Removes the WinVNC service and continues reading the command−line.
−settings
Tells a running copy of WinVNC to show the User Properties box.
−defaultsettings
Tells a running copy of WinVNC to show the Default Properties box.
−connect host
Tells a running copy of WinVNC to initiate an outgoing connection to a listening
viewer running on the specified machine. This is the equivalent of the 'Add New
Client' menu option. You can put multiple −connect options on one command
line to connect to multiple viewers at once.
−kill
Kills a running copy of WinVNC.
−about
Tells a running copy of WinVNC to show its About box.
If no options are given then WinVNC runs normally. Multiple options may be given, so,
for example, to upgrade a currently running WinVNC service to a new version, you
could use:
WinVNC_new −remove −install
which will stop & remove the old copy & install the new one as a service, or
WinVNC_new −kill −run
which will stop the running copy & run the new version as an application.
WinVNC − Advanced Settings
Extra options have been added to WinVNC for use primarily by system administrators,
to tailor the server's behaviour to meet their particular needs. The options are
DWORD values which can be set in the system registry, and tools such as the Windows
Policy Editor can be used to apply these settings across a large number of machines.
Versions 3.3.2 R5 and later use a more sophisticated organisation of these options to
allow more flexibility. It also makes it rather complex, so we're thinking about
alternative ways of doing this. WinVNC will currently look for settings in the following
places:
1. Local machine−specific settings. Options specified here are not overridable.
Location:
HKEY_LOCAL_MACHINE\Software\ORL\WinVNC3\
2. Local default user settings. Location:
HKEY_LOCAL_MACHINE\Software\ORL\WinVNC3\Default
3. Local per−user settings. These override the local default user settings. If there
is no current user, the username SYSTEM will be used. Location:
HKEY_LOCAL_MACHINE\Software\ORL\WinVNC3\<username>
4. Global per−user settings. These are only read if AllowProperties has not
been set to zero (see below) Location:
HKEY_CURRENT_USER\Software\ORL\WinVNC3 Most options can only be
specified in a subset of these places, as specified in each option's description below.
Advanced Options:
AuthRequired
By default, all WinVNC servers will not accept incoming connections unless the
server has had its password field set to a non−null value. This restriction was
placed to ensure that misconfigured servers would not open security loopholes
without the user realising. If a server is only to be used on a secure LAN,
however, it may be desirable to forego such checking and allow machines to
have a null password. Setting this registry value to zero will disable
null−password checking by WinVNC. Local machine−specific setting.
AllowLoopback
By default, WinVNC servers disallow any vncviewer connections from the same
machine. For testing purposes, or, potentially, when using multiple instances
of WinVNC on Windows Terminal Server, this behaviour is undesirable.
Setting this registry entry to 1 will cause local−loopback connections to be
allowed. Setting it to zero will filter out such connections. Local
machine−specific setting.
AllowProperties
If this is set to zero, the user is not allowed to view the properties dialog and
hence cannot change any settings, including the password. Note that this stops
all global per−user settings. A valid password must therefore be in force before
using this setting, generally in the local default−user setting. Local per−user
setting.
AllowShutdown
If this is set to zero, the user is not allowed to close down WinVNC. Local
per−user setting.
AuthHosts
The AuthHosts setting is, unlike the other settings, a REG_SZ string. It is used to
specify a set of IP address templates which incoming connections must match in
order to be accepted. By default, the template is empty and connections from all
hosts are accepted. The template is of the form:
+[ip−address−template]
?[ip−address−template]
−[ip−address−template]
In the above, [ip−address−template] represents the leftmost bytes of the desired
stringified IP−address. For example, +158.97 would match both 158.97.12.10
and 158.97.14.2. Multiple match terms may be specified, delimited by the ":"
character. Terms appearing later in the template take precedence over earlier
ones. e.g. −:+158.97: would filter out all incoming connections except those
beginning with 158.97. Terms beginning with the "?" character are treated by
default as indicating hosts from whom connections must be accepted at the
server side via a dialog box. The QuerySetting option determines the precise
behaviour of the three AuthHosts options. Local machine−specific setting.
QuerySetting
The QuerySetting allows individual users to tailor the degree of paranoia
expressed by the per−machine AuthHosts setting. It is a DWORD value ranging
from zero (maximum availability) to four (maximum security). The following
table indicates how the value affects the AuthHosts behaviour:
0 − +:Accept, ?:Accept, −:Query
1 − +:Accept, ?:Accept, −:Reject
2 − +:Accept, ?:Query, −:Reject [Default]
3 − +:Query, ?:Query, −:Reject
4 − +:Query, ?:Reject, −:Reject
By default, value 2 (Obey AuthHosts) will be assumed. Local global per−user
setting.
QueryTimeout
The QueryTimeout setting indicates the number of seconds for which the Accept
Connection dialog (see AuthHosts and QuerySetting options) will be displayed
before rejecting the incoming connection automatically. Local global per−user
setting.
AutoPortSelect
Causes WinVNC to select the first available display number automatically.
Corresponds to the 'Auto' checkbox in the Properties dialog. Local or Global
per−user setting
CORBAConnect
Only relevant in internal AT&T version. Local or Global per−user setting
DebugLevel
DebugMode
Run−time logging of all internal debug messages is now supported. Log data
may be output to a file or a console window (or the MSVC debugger if the
program was compiled with debugging active.) Two registry keys are used:
DebugMode indicates which logging methods to use,
[1 = MSVC debugger]
2 = Output to log file Winvnc.log in the WinVNC directory
4 = Output to a console window, displayed on−screen
Any combination of the above values may be used. e.g. DebugMode=6 will
cause output to be sent both to the WinVNC.log file and to the a console window
on the desktop.
DebugLevel indicates how much debug information to present. Any positive
integer is valid. Zero indicates that no debugging information should be
produced and is the default. A value of around 10−12 will cause full debugging
output to be produced. Local machine−specific setting.
ConnectPriority
By default, all WinVNC servers will disconnect any existing connections when
an incoming, non−shared connection is authenticated. This behaviour is
undesirable when the server machine is being used as a shared workstation by
several users or when remoting a single display to multiple clients for vewing, as
in a classroom situation.
ConnectPriority indicates what WinVNC should do when a non−shared
connection is received:
0 = Disconnect all existing connections.
1 = Don't disconnect any existing connections.
2 = Refuse the new connection.
This is a Local machine−specific setting.
IdleTimeout
This setting tells WinVNC how many seconds a connected VNC client may
remain idle for (no input events or update requests) before being disconnected. If
this setting is not specified or is set to zero then no timeout is enforced. Local or
Global per−user setting
InputsEnabled
Corresponds (inversely) to the 'Disable Remote keyboard and pointer' option in
the Properties dialog box. Local or Global per−user setting
LockSetting
WinVNC can be made to take actions when a viewer disconnectsby setting this
value as follows:
0 − none
1 − lock workstation on disconnect (not currently implemented)
2 − logoff on disconnect
Local or Global per−user setting
LoopbackOnly
By default, WinVNC servers accept incoming connections on any network
adapter address, since this is the easiest way of coping with multihomed
machines. In some cases, it is preferable to listen only for connections
originating from the local machine and aimed at the "localhost" adapter − a
particular example is the use of VNC over SSH to provide secure VNC. Setting
this registry entry to 1 will cause WinVNC to only accept local connections −
this overrides the AllowLoopback and AuthHosts settings. Setting this entry to
zero causes WinVNC to accept connections on any adapter and is the default
setting.Local machine−specific setting.
Password
PollUnderCursor, PollForeground, PollFullScreen, OnlyPollConsole,
OnlyPollOnEvent
These correspond to the options in the Properties dialog box. Local or Global
per−user settings
PortNumber
specifies the port number to be used for VNC. You will need to disable
AutoPortSelect to use this.
RemoveWallpaper
Indicates whether or not WinVNC should remove the user's background
wallpaper when an incoming connection is made. It is necessary to reconnect in
order for this setting to take effect. Local or Global per−user setting
SocketConnect
This corresponds to the 'Accept Socket Connections' option in the properties
dialog box and is a Local or Global per−user setting.
VNCHooks − Advanced Settings
WinVNC uses a special library, VNCHooks, to hook into the other running applications
and retrieve notifications of areas of the screen being changed. The VNCHooks library
uses the messages sent to visible Windows to decide which areas need considering for
update. Not all applications use the same method of updating the screen, so you can
tweak the method used by WinVNC for particular applications by editing the registry.
All the entries listed can be found under
HKEY_CURRENT_USER\Software\ORL\VNCHooks\Application_Prefs
• use_GetUpdateRect
When a window recieves a message, (WM_PAINT), indicating that it should repaint
itself, it is possible to find out precisely which regions have changed, so that WinVNC
need only scan those for potential updates,increasing efficiency. However, this can cause
graphical glitches occasionally, particularly when an application scrolls the contents of
its window, in which case only the revealed section of the window is marked as needing
to be updated. If these glitches prove to be a problem then edit the
<appname>\use_GetUpdateRect entry in the registry. A value of one indicates
that this optimisation will be used, while a value of zero indicates that it will not.
• use_Timer
A number of Windows applications, most notably the Clock program, use WM_TIMER
events to trigger updates to their displays, rather than WM_PAINT messages. By default,
timer messages are not used to notify WinVNC of potential updates, since many
programs use timer events for purposes other than updating the screen. As a result, the
clock and a few other applications don't normally update correctly under WinVNC. The
fix to this is to edit the <appname>\use_Timer entry in the registry. A value of one
indicates that WM_TIMER messages will trigger WinVNC updates, while a value of
zero indicates that they will not.
• use_KeyPress
Some Windows applications write characters directly to the screen when a user types into
a window, rather than using WM_PAINT messages to cause the text to be redrawn. To
fix this, WinVNC can scan the window every time a key is pressed, in order to catch the
change. To set this value for a problem application, edit the
<appname>\use_KeyPress entry in the registry. A value of one indicates that key
presses will cause updates, while a value of zero indicates that they will not.
• use_LButtonUp, use_MButtonUp, use_RButtonUp,
Some Windows applications update the display directly in response to mouse clicks,
without using intermediate WM_PAINT messages, for example. In order to catch such
updates, it is necessary to trigger WinVNC to update the relevant window whenever the
left mouse button is released. To set this value for a problem application, edit the
<appname>\use_LButtonUp entry in the registry. A value of one indicates that
left−button clicks will cause updates, while a value of zero indicates that they will not.
The same rules apply to the middle and right buttons using the appropriate value name.
• use_Deferral
The VNCHooks library catches messages sent to windows before they are dealt with by
the window. As a result, sending an update message to WinVNC to indicate the potential
change can result in WinVNC sending the updated area to the client before it has actually
been redrawn by the application! This is a common problem, especially on
multiprocessor versions of NT, so deferred updates are used by default. Deferred updates
are handled by posting a custom message back into the window's own message queue
rather than posting to WinVNC directly. By the time this custom message is seen again
by the VNCHooks library, the message that caused it will have been handled and the
update can then be forwarded to WinVNC without danger of being handled prematurely.
A few programs don't handle these extra messages in their queue very well, so this
optimisation is optional. It can be set by editing the <appname>\use_Deferral
entry in the registry. A value of one indicates that deferred updates will be used, while a
value of zero indicates that they will not.
Running on other Win32 systems
WinVNC runs fine on NT3.51 but the absence of a system tray means that the Properties
dialog cannot be accessed. In addition, Ctrl−Alt−Del from clients cannot be correctly
interpreted under NT 3.51, limiting WinVNC's usefulness when run as a service on this
platform. It also runs on NT5 beta. If you have a choice we recommend NT4.0 with the
latest service packs installed.
Problems?
If you have difficulties which are not covered by this document, try reading the FAQ. If
that doesn't help then try the mailing list.
If you try to contact the developers directly, please remember that VNC has hundreds of
thousands of users, and we cannot, in general, respond to individual queries. We will
read your message, but don't expect an answer!
VNCviewer for Windows
If you install the Windows server, the viewer is also installed and is available from the
VNC section of the start menu. But you can start it in other ways, and you can use the
viewer without installing the server.
You can run the windows vncviewer from the command line or from a shortcut and it
will prompt you for a display:
vncviewer
You can specify a display on the command line:
vncviewer snoopy:2
And you can run it with −h to get a list of other important options. The full list is below.
These can all take either − or / as the switch character. Most of the options can also
be set from the 'Options...' dialog box which is available from the initial connection
prompt before connecting, and some from the system menu by clicking the VNC logo
in the top−left corner of the window after connection, and selecting 'Connection
options...'.
The system menu also allows you to see some information about the connection, start
new connections, and send a Ctrl−Alt−Del to a remote machine. This will only have
an effect if the remote server is able to interpret it, currently only true for WinVNC
running as a service under NT 4.
Keystrokes such as Ctrl−Esc and Alt−Tab may be interpreted at the local (viewer)
machine. If you want to send them to the remote machine, you can use the options on
the viewer menu to send individual Ctrl−down, Ctrl−up, Alt−down and Alt−up
keystrokes. For example, to type Ctrl−Esc on the remote machine, send Ctrl−down
using the menu, press Esc, and then send Ctrl−up (or just tap the Ctrl key) to release the
Ctrl key at the remote end.
Command line options:
−shared
When you make a connection to a VNC server, all other existing connections
are normally closed. This is for security reasons, and because we normally
think of VNC as a tool for mobility: your desktop follows you from place to
place. This option asks the server to leave any existing connections open,
allowing you to share the desktop with someone already using it. Some servers
have options to change the default behaviour and to override this request.
−8bit
The viewer will normally accept whatever pixel format the server offers and do
the translation locally. This forces it to request 8−bit true−colour (BGR233)
from the server, which will reduce network traffic. Useful over modems.
−config file
You can save all the details of an open connection to a file using a command
from the menu. You can then restart that connection at a later date by
specifying the name of the file using this switch.
−register
This tells the Windows shell that .vnc files are associated with the vncviewer.
You should then be able to double−click on them to start the session.
Sometimes Windows seems to need restarting before this take effect.
−scale n/m
Specify a scaling factor for the local display. The values n and m should be
integers. The '/' and the m can be omitted if m=1.
−emulate3
Users with a two−button mouse can emulate a middle button by pressing both
buttons at once if this option is enabled on the command line or in the dialog
box. Note: On recent versions of the viewer this is the default, so there's now a
−noemulate3 option to turn it off if wanted.
−noemulate3
Opposite of −emulate3
−swapmouse
This option was more commonly used before the 3−button emulation was
available. Normally the PC buttons left−middle−right are mapped on to X
buttons 1,2,3. This switch causes them to be mapped onto buttons 1,3,2, which
may be more useful for two−button users who only have left−right, because
they will then get buttons 1 & 2 instead of 1 & 3.
If combined with 3−button emulation, this also causes the middle button to
emulate button 3 instead of button 2. This may be useful if you use button 2
more.
−emulate3timeout
When using 3−button emulation, both mouse buttons must be pressed within a
certain period for them to be registered as a single middle−click instead of
separate left and right clicks. This option allows that time period to be specified
in msec. The default is 100.
−emulate3fuzz
When using 3−button emulation, both mouse buttons must be pressed within a
certain distance of each other for them to be registered as a single middle−click.
This option allows that distance to be specified in pixels. The default is 4.
−fullscreen
This causes connections to start in full−screen mode by default. See below for
more details.
−listen
In the internal version of VNC used at AT&T Labs
Cambridge, the server can initiate connections to the
clients under CORBA control. This switch puts
vncviewer into listening mode where it can accept these connections, but it also
has a useful side−effect which may be of interest to those outside AT&T using
the public version. A listening vncviewer does not pop up a connection dialog,
but instead installs itself in the system tray. From there you can easily start up
new connections and can set default options to be used for them during this
instance of the program. RECENT NEWS! The latest versions of WinVNC
can initiate the connection to a viewer using the 'Add New Client' menu option.
For this to work, the viewer must be in listening mode.
−disableclipboard
Clipboard changes caused by cutting or copying at either the viewer or server
end are normally transmitted to the other end. This option disables clipboard
transfers.
−belldeiconify
VNC allows for the transmission of a 'bell' character, causing a beep at the
viewer if it has sound facilities. You can set the sound to be used for the bell
under the VNCviewer section of 'Sounds' in the Control Panel. Often a beep
will happen because you are being notified of something such as email arriving
or compilation finishing. This switch causes a minimized vncviewer to be
un−minimized when a bell character is received.
−nocursor, −dotcursor, −normalcursor
Most VNC servers send their cursor as part of the screen image that is
displayed in the viewer. Having a local cursor in addition to this can be
distracting. The default is for the viewer to use a small dot to show the position
of the local cursor, and this is our recommended mode of use. You can use the
−nocursor option to turn off this local cursor completely, or −normalcursor to
leave it at the default Windows 'arrow'. Some things to note here:
◊ When you press a mouse button, it is the local mouse position that is
used to send the event. On a slow network, the remote cursor may lag
behind the local one a bit. You don't need to wait for it to catch up
before you click, but if you have switched off the local cursor display,
it can be harder to know exactly where you're clicking!
◊ The X−based server has an option which tells it not to show a cursor.
This can be useful if combined with −normalcursor at the viewer,
particularly on slow networks. However, the cursor will then never
change shape − it will always be the arrow.
We like the default dot the best!
−keyboard kbdname
Windows uses an internal and not very helpful name for the keyboard layout
currently selected for an application. You can see the one being used by
vncviewer if you select 'Connection Info' from the system menu of the viewer
window. If you change the keyboard settings and then make a note of this,
you can specify it on the command line to cause vncviewer to attempt to load
this in the future. Note that vncviewer does not currently support 'dead keys',
and that the differences between language and keyboard are confusing and the
way they are handled is different in Windows 95 and NT. But this may help a
bit.
−logfile filename
VNCviewer (R6 and later) has a logging mechanism which can save some
debugging information to a file or display it on a console. This option specifies
the name of a file to which a log will be written.
−loglevel n
This option controls the amount of logging information sent to the log file.
The default is zero, and higher values (up to about 12) will provide more detail.
−console
In addition to, or instead of, logging to a file, this option will cause the
debugging information to be sent to a console window.
−viewonly
In View−only mode, no mouse or keyboard events will be sent back to the
server. This is useful for teaching sessions or other situations where you want to
observe but don't want to interfere.
−restricted
In restricted mode, most of the items are removed from the menu, so that the
user cannot, for example, send a Ctrl−Alt−Del to the remote end.
Full−screen mode
Vncviewer can now be switched into a fullscreen mode. This is particularly useful
when connecting to a remote screen which is the same size as your local one. If the
remote screen is bigger, you won't get any scrollbars, but you can scroll by bumping the
mouse against the edge of the screen.
To leave fullscreen mode you must disable it from the menu, but the menu is no longer
visible! So you have to bring the taskbar to the front by typing Ctrl−Esc Esc, and then
right−click on the vncviewer icon. A dialog box will appear when you select
fullscreen mode to remind you of this; if, after a while you get annoyed with the dialog
box, you can disable it by creating a DWORD registry value named
HKEY_CURRENT_USER\Software\ORL\
VNCviewer\Settings\SkipFullScreenPrompt
and setting it to 1. A simpler method will be in a future version!
You can cause new connections to start in fullscreen mode using the −fullscreen
command−line option.
See also 'What's new in the Windows VNC package?'
VNCviewer for X (3.3.2r3 and earlier)
Note: this page documents the Xlib−based viewer in the 3.3.2r3 and earlier releases.
For the current release, see the documentation for the new Xt−based viewer.
Run it and give the display as an argument:
vncviewer snoopy:2
where 'snoopy' is the name of the machine, and '2' is the display number of the VNC
server on that machine. The −h argument will show you the other options. The
important ones are as follows:
−shared
When you make a connection to a VNC server, all other
existing connections are normally closed. This option
requests that they be left open, allowing you to share the
desktop with someone already using it.
−display Xdisplay
This allows you to specify the X display on which the
VNCviewer window should appear.
−passwd password−file
If you are on a filesystem which gives you access to the
password file used by the server, you can specify it here to
avoid typing it in.
−viewonly
Specifies that no keyboard or mouse events should be sent to
the server. Useful if you want to view a desktop without
interfering; often needs to be combined with −shared.
−geometry geometry
Standard X position and sizing specification.
−bgr233
This tells the VNC server to send pixels which are only 8 bits
deep. If your server desktop is deeper than this then it will
translate the pixels before sending them. Less data will
generally be sent over then network, which can be a big
advantage on slow links, but the server may have to work a bit
harder, and you may get some colour mismatches. This is the
8−bit true colour pixel format used by the java client, with the
most significant two bits of each byte representing the blue
component, the next three bits representing green and the least
significant three representing red, hence 'bgr233'.
−raw, −copyrect, −rre, −corre, −hextile,
−nocopyrect, −norre, −nocorre, −nohextile
These options affect which encodings vncviewer tells the VNC
server it can cope with. Normally it requests CopyRect,
Hextile, CoRRE and RRE in that order. The options alter this
behaviour in the obvious way, e.g. specifying just −raw means
Raw will be requested before any of the others, specifying
−norre means don't request RRE, etc.
−depth d
This is only useful on a (real) X server which supports multiple
depths. On such a display vncviewer will try to find a Visual
of the given depth. If successful this means that the
appropriate pixel format will be requested from the VNC
server. You cannot use this to force a particular depth from
the VNC server. The only option which does this is
−bgr233.
−truecolour
vncviewer will try to find an X visual of the TrueColor class.
−owncmap
vncviewer will try to find a PseudoColor visual and use its own
Colormap.
−period ms
This tells vncviewer not to request incremental framebuffer
updates more often than the given period in milliseconds. If
you have a very fast client and server, you can use this option
to limit the rate of updates − this can result in using less
network bandwidth.
−wmdecoration wxh
Normally the viewer window will be the same size as the
desktop you're connecting to, plus any decorations added by
your local window manager. If your local screen is too small to
display this then it will make the window as big as it can while
still allowing room for decorations, and display scrollbars.
This option allows you to specify how big your window
manager's decorations are, so that it can make this decision
correctly. In particular, some people run without a local
window manager, and specifying −wmdecoration 0x0 will then
allow a completely full−screen window.
−rawdelay ms
This is useful for checking exactly which parts of the screen
are being updated. For each update rectangle vncviewer puts
up a black rectangle for the given time before putting up the
pixel data. This only highlights pixel data sent using the raw
encoding.
−copyrectdelay ms
This works as with −rawdelay above, but highlights the areas
copied using the copyrect encoding.
−debug
This prints out all the data received from the VNC server in
both hex and ASCII.
−listen
This is used for AT&T's internal version of VNC. It causes
vncviewer to listen on port 5500+<display−number> for
reverse connections from a VNC server. See
http://www.uk.research.att.com/vnc/internalversion.html
X−based VNC server
Make sure you've read 'Getting Started' for introductory information.
Xvnc is the Unix VNC server, which is based on a standard X server. Applications
can display themselves on it as if it were a normal X display, but they will actually
appear on any connected VNC viewers rather than on a physical screen.
So Xvnc is really two servers in one. To the applications it is an X server, and to the
remote VNC users it is a VNC server. By convention we have arranged that the VNC
server display number will be the same as the X server display number, which means
you can use eg. snoopy:2 to refer to display 2 on machine 'snoopy' in both the X
world and the VNC world.
Normally you will start Xvnc using the vncserver script, which is designed to
simplify the process, and which is written in Perl. You will probably want to edit this to
suit your preferences and local conditions. We recommend using vncserver rather
than running Xvnc directly, but Xvnc has essentially the same options as a standard X
server, with a few extensions. Running Xvnc −h will display a list.
As mentioned in Getting Started , vncserver can be run with no options at all. In this
case it will choose the first available display number, start Xvnc as that display, and run
a couple of basic applications to get you started. You can also specify the display
number, in which case it will use that number if it is available and exit if not, eg:
vncserver :13
Run with the −help argument to see the other options. The important ones are as
follows:
−name name
Each desktop has a name which may be displayed by the
viewer. It defaults to 'X' but you can change it with this option.
−geometry widthxheight
Specify the size of the desktop to be created. Default is
1024x768.
−depth depth
Specify the pixel depth in bits of the desktop to be created.
Default is 8.
−pixelformat format
Specify pixel format for server to use (BGRnnn or RGBnnn)
In general, you can specify standard X server arguments to vncserver, and they will
be passed through to Xvnc. Again, Xvnc −help will list its options, including:.
−inetd
This significantly changes Xvnc's behaviour so that it can be
launched from inetd. See the information on the extras page for
details.
−alwaysshared
Always treat new clients as shared (i.e. ignore client's shared
flag).
−nevershared
Never treat new clients as shared (i.e. ignore client's shared
flag).
−dontdisconnect
Don't disconnect existing clients when a new "non−shared"
connection comes in. Instead the new connection is refused.
New "shared" connections are still allowed in the normal way.
−localhost
Only allow connections from the same machine. Useful if you
use SSH and want to stop non−SSH connections from any
other hosts. See the guide to using VNC with SSH
−cc n
Sets the colour Visual class used by the server. Some X
applications don't cope too well with the TrueColor visual
normally used by an 8−bit−deep Xvnc. You can make the
server use a PseudoColor visual by specifying −cc 3.
−economictranslate
The server normally uses a lookup table for translating pixel
values when the viewer requests a different format from the
native one used by the server. This can use up to 256Kbytes
per connected viewer, so if you have many viewers you may
wish to specify this option which will save memory at the
expense of a little bit of speed. Only relevant for 16−bit−deep
desktops.
−deferupdate n
Xvnc now uses a "deferred update" mechanism which
enhances performance in many cases. After any change to the
framebuffer, Xvnc waits for this number of milliseconds
(default 40) before sending an update to any waiting clients.
This means that more changes tend to get coalesced together in
a single update. Setting it to 0 results in the same behaviour as
previous versions of Xvnc.
The script ~/.vnc/xstartup is executed after the server starts. If you want to change the
window manager used in your VNC desktop, for example, this is where you should do
it.
The server also writes log files in the ~/.vnc directory. These can be useful for
tracking down configuration problems and startup errors.
Reverse connections
Xvnc can now make reverse connections to a listening viewer (normally connections
are made the other way round − the viewer connects to the server). This is done with a
helper program called vncconnect. Simply run:
vncconnect host
where host is the host where the listening viewer is running. If it's not listening on the
default port of 5500, you can specify host:port instead. Vncconnect is just a simple X
program which sets a property on the root window − it is Xvnc which actually makes
the connection. Note that the DISPLAY environment variable must be set as
appropriate to control Xvnc − if you run it on an ordinary X server then nothing will
happen. Also note that any feedback about the success or failure of the connection will
appear in the log file for Xvnc rather than on the terminal where you run vncconnect.
X VNC History
The version and release numbers are explained on the main VNC history page. This page
primarily concerns changes to the Unix and Java packages. The changes to the Windows
packages are listed separately here.
The important changes in each version are summarised here; see the source package for
more details.
Changes in version 3.3.3R2
vncconnect
• New helper program for Xvnc, which causes it to make reverse connections to a
listening VNC viewer. This is similar to WinVNC's "Add New Client" /
−connect feature.
Xvnc
• New option −inetd incorporating Andre Moreira's iXvnc functionality.
• Fix to hextile encoding, thanks to Michael A. Fetterman − this may be the cause
of the "0 bpp problem".
• New connect out feature by using the vncconnect command.
• Fix some non−ascii capitalisation issues, thanks to Vlad Harchev.
• Fix bounding box calculation for arcs.
• Other minor bug fixes.
vncserver
• Sets DISPLAY to use unix domain socket if possible (:dpynum rather than
host:dpynum).
• Option −kill should now work if you give it $DISPLAY as an argument.
vncviewer
• Fix Xaw string copy bug, thanks to Tim Waugh − this fixes the problems with
the password dialog with recent versions of linux
Java viewer
• Improve keyboard handling for Latin−1 characters and ctrl−_.
• Change constants so (hopefully) will compile with JDK 1.3
Changes in version 3.3.3R1
Xvnc
• Xvnc now uses a "deferred update" mechanism which enhances performance in
many cases. After any change to the framebuffer, Xvnc waits for a time (default
40ms) before sending an update to any waiting clients. This means that more
changes tend to get coalesced together in a single update. The time can be
changed with the −deferupdate option. Setting it to 0 results in the same
behaviour as previous versions of Xvnc.
• There are a couple of security enhancements. The −localhost option now causes
Xvnc to only listen on the loopback interface. This means that it shouldn't be
vulnerable to port scanning. Also a "too many tries" feature on authentication
failure is now implemented. After five failed attempts, further attempts are
delayed by an exponentially increasing amount, starting at ten seconds.
vncserver
• Uses SO_REUSEADDR when testing to see which ports are free.
vncviewer
• When given −owncmap in full screen mode, tries to install the colormap. This
works only when no window manager is running.
• The popup window now always uses the default colormap.
• The "backing store" attribute can now be set on the desktop window via the
"backingStore" resource. It defaults to "Always", meaning the X server should
keep the contents of the viewer window when it gets obscured by other
windows.
Release date: 9 February 2000
Changes in version 3.3.3
Xvnc
New options −alwaysshared, −nevershared, −dontdisconnect and −localhost.
Fixed problem with 16−bit desktops on Solaris and other platforms.
Default dpi now 75.
Don't die on non−existent font directory at startup (common RedHat 6 problem).
Copes with up to 5 pointer buttons.
Fixed zero−area copy bug.
Fixed occasional bug when multiple clients and one disconnects.
vncserver
Work around bug in perl 5.005_02 which caused bad display names such as "snoopy:".
vncpasswd
Fixed minor bugs.
vncviewer
Completely new viewer based on Xt. Development has ceased on the old Xlib−based
viewer.
Release date: 19 August 1999
Changes in version 3.3.2 R3
Xvnc
Fixed an important memory leak.
vncserver
The vncserver script should cope better with different Perl versions.
X cookies generated more securely.
vncviewer
Uses shared memory on Linux.
More secure clearing of password.
Release date: 14 December 1998
Changes in version 3.3.2 R2
Xvnc
Cutting text inside Xvnc works to a remote viewer again.
Resource leak fixed which limited total number of times Xvnc could be
connected to.
vncserver
Warns of likely failure when starting a 16 bpp desktop on Solaris.
No longer tries to find out the geometry, depth and pixel format of the current X
display being used. This feature can be enabled if desired by uncommenting a
line in the script.
Release date: 23 July 1998
Changes in version 3.3.2
The changes in 3.3.2 will have little effect on the normal user. There should be a slight
increase in speed in certain circumstances, but the majority of changes are 'under the
hood'. The X server is now based on XFree86 3.3.2. This should make it easier to
build on more platforms, and means that it is now based on X11R6.3 rather than
X11R6.
X vncviewer
• −viewonly, −geometry, −wmdecoration options
• Options for debugging servers: change −delay to −rawdelay, plus
−copyrectdelay, −region
• Options allowed before or after host:dpynum argument
• Handle changes to local X keyboard mapping
• Rudimentary single bit−per−pixel support (thanks to Chris Hooper)
Xvnc server
• Based on XFree86 3.3.2 distribution
• Faster pixel−format translation routines
• Improved HTTP server. Now performs substitution on any files ending in .vnc,
so that default options to the VNC applet can be configured on the server side.
• −economictranslate option to use less memory−hungry translation
• Supports viewers requesting only a subregion of the screen.
• Better log file messages with timestamps.
VNCviewer for X (3.3.3 onwards)
Note: this page documents the Xt−based viewer in the 3.3.3 release. For previous
releases, see the documentation for the old Xlib−based viewer.
If you just run the viewer from the command line, it will prompt you for a VNC server to
connect to:
vncviewer
Alternatively, specify the VNC server as an argument, e.g.:
vncviewer snoopy:2
where 'snoopy' is the name of the machine, and '2' is the display number of the VNC
server on that machine. Either the machine name or display number can be omitted. So
for example ":1" means display number 1 on the same machine, and "snoopy" means
"snoopy:0" i.e. display 0 on machine "snoopy".
If the VNC server is successfully contacted, you will be prompted for a password to
authenticate you. If the password is correct, a window will appear showing the desktop
of the VNC server.
Popup window
The viewer has a popup window containing a set of buttons which perform various
actions. It is usually brought up by pressing F8, but this is customisable, as is the entire
contents of the popup. Actions which buttons in the popup window can perform include:
• switching in and out of full−screen mode
• quitting the viewer
• generating arbitrary key and mouse events, e.g. sending ctrl−alt−del
• transferring the clipboard to or from the VNC server
By default, key presses in the popup window get sent to the VNC server and dismiss the
popup. So to get an F8 through to the VNC server simply press it twice.
See the section on customisation below for how to customise the contents of the popup
window.
Full screen mode
A full−screen mode is supported. This is particularly useful when connecting to a remote
screen which is the same size as your local one. If the remote screen is bigger, you can
scroll by bumping the mouse against the edge of the screen.
Unfortunately this mode doesn't work completely with all window managers, since it
breaks all the X window management conventions. It tends to work better when the
viewer is started in full−screen mode than when switching to it from normal mode.
Command line options
You can get a list of options by giving −h as an option to vncviewer. Most of these
options can also be specified as X resources − see the section on customisation below.
−shared
When you make a connection to a VNC server, all other existing connections are
normally closed. This option requests that they be left open, allowing you to
share the desktop with someone already using it.
−display Xdisplay
Specifies the X display on which the VNC viewer window should appear.
−passwd password−file
If you are on a filesystem which gives you access to the password file used by
the server, you can specify it here to avoid typing it in. It will usually be
"~/.vnc/passwd".
−viewonly
Specifies that no keyboard or mouse events should be sent to the server. Useful
if you want to view a desktop without interfering; often needs to be combined
with −shared.
−fullscreen
Start in full−screen mode.
−geometry geometry
Standard X position and sizing specification.
−bgr233
Tells the VNC server to send pixels which are only 8 bits deep. If your server
desktop is deeper than this then it will translate the pixels before sending them.
Less data will generally be sent over then network, which can be a big advantage
on slow links, but the server may have to work a bit harder, and you may get
some colour mismatches. "BGR233" means an 8−bit true colour pixel format,
with the most significant two bits of each byte representing the blue component,
the next three bits representing green and the least significant three representing
red. This format is also used by the java client.
−encodings encodings
This option specifies a list of encodings to use in order of preference, separated
by spaces. The default is "copyrect hextile corre rre", or "raw copyrect hextile
corre rre" for a VNC server on the same machine. For example, to use only raw
and CopyRect, specify "raw copyrect".
−owncmap
Try to use a PseudoColor visual and a private colormap − this allows the VNC
server to control the colormap.
−truecolour
Try to use a TrueColor visual.
−depth d
This is only useful on a (real) X server which supports multiple TrueColor
depths. On such a display vncviewer will try to find a Visual of the given depth.
If successful this means that the appropriate pixel format will be requested from
the VNC server. You cannot use this to force a particular depth from the VNC
server. The only option which does this is −bgr233.
−listen
Causes vncviewer to listen on port 5500+<display−number> for reverse
connections from a VNC server. WinVNC supports reverse connections initiated
using the 'Add New Client' menu option or the '−connect' command−line option.
Xvnc now supports reverse connections with a helper program called
vncconnect. It is also used for our internal version of VNC − see
http://www.uk.research.att.com/vnc/internalversion.html.
Customisation and X resources
The behaviour of vncviewer is extremely customisable using X resources. You can set X
resources by any of the usual means − in an app−defaults file such as .Xresources, or on
the command line with the '−xrm' option. See the X window system documentation for
details.
The application resources are:
shareDesktop (option −shared)
Whether to leave other viewers connected. Default false.
viewOnly (option −viewonly)
Block mouse and keyboard events. Default false.
fullScreen (option −fullscreen)
Full screen mode. Default false.
passwordFile (option −passwd)
File from which to get the password (as generated by the vncpasswd program).
Default is null, i.e. to request password from the user.
passwordDialog
Whether to use a dialog box to get the password (true) or get it from the tty
(false). Irrelevant if passwordFile is set. Default false.
encodings (option −encodings)
A list of encodings to use in order of preference, separated by spaces. Default is
null, which actually means "copyrect hextile corre rre", or "raw copyrect hextile
corre rre" for a VNC server on the same machine.
useBGR233 (option −bgr233)
Always use the BGR233 (8−bit) pixel format on the wire, regardless of the
visual. Default is false (though BGR233 is used anyway for non−TrueColor
visuals with forceOwnCmap false).
nColours
When using BGR233, try to allocate this many "exact" colours from the
BGR233 colour cube. When using a shared colormap, setting this resource lower
leaves more colours for other X clients. Irrelevant when using truecolour.
Default is 256 (i.e. all of them).
useSharedColours
If the number of "exact" BGR233 colours successfully allocated is less than 256
then the rest are filled in using the "nearest" colours available. This resource
says whether to only use the "exact" BGR233 colours for this purpose, or
whether to use other clients' "shared" colours as well. Default true (i.e. use other
clients' colours).
forceOwnCmap (option −owncmap)
Try to use a PseudoColor visual and a private colormap − this allows the VNC
server to control the colormap. Default false.
forceTrueColour (option −truecolour)
Try to use a TrueColor visual. Default false.
requestedDepth (option −depth)
If forceTrueColour is true, try to use a visual of this depth. Default 0 (i.e. any
depth).
useSharedMemory
Whether to use the MIT shared memory extension if on the same machine as the
X server. Default true.
wmDecorationWidth
wmDecorationHeight
The total width and height taken up by window manager decorations. This is
used to calculate the maximum size of the VNC viewer window. Default is
width 4, height 24.
bumpScrollTime
bumpScrollPixels
When in full screen mode and the VNC desktop is bigger than the X display,
scrolling happens whenever the mouse hits the edge of the screen. The
maximum speed of scrolling is bumpScrollPixels pixels every bumpScrollTime
milliseconds. The actual speed of scrolling will be slower than this, of course,
depending on how fast your machine is. Default 20 pixels every 25 milliseconds.
popupButtonCount
The number of buttons in the popup window. See below for how to customise
the buttons.
rawDelay
This is useful for debugging VNC servers by checking exactly which parts of the
screen are being updated. For each update rectangle vncviewer puts up a black
rectangle for the given time in milliseconds before putting up the pixel data.
This only highlights pixel data sent using the raw encoding. Default 0 (i.e. don't
do it).
copyRectDelay
Similar to rawDelay, but highlights the areas copied using the copyrect
encoding.
How to customise the popup window
Set the number of buttons with the popupButtonCount resource, e.g.:
*popupButtonCount: 2
For each button, set the label, and override the button press translations, e.g.:
*popup*button1.label: Left mouse button click at 100,100
*popup*button1.translations: #override\n\
<Btn1Down>,<Btn1Up>: SendRFBEvent(ptr,100,100,1)\
SendRFBEvent(ptr,100,100,0)
*popup*button2.label: Send "Think thin!"
*popup*button2.translations: #override\n\
<Btn1Down>,<Btn1Up>:\
SendRFBEvent(key,T) SendRFBEvent(key,h)\
SendRFBEvent(key,i) SendRFBEvent(key,n)\
SendRFBEvent(key,k) SendRFBEvent(key,space)\
SendRFBEvent(key,t) SendRFBEvent(key,h)\
SendRFBEvent(key,i) SendRFBEvent(key,n)\
SendRFBEvent(key,exclam)
How to customise the desktop window
You can override translations on the desktop window. For example to change the key
used to bring up to popup window from F8 to Escape, and make F12 switch in and out of
full screen mode:
*desktop.translations: #override\n\
<Key>F8: SendRFBEvent()\n\
<Key>Escape: ShowPopup()\n\
<Key>F12: ToggleFullScreen()
Actions
These are the actions which you can use in translations:
ShowPopup()
HidePopup()
Show and hide the popup window, respectively.
SendRFBEvent()
Send an RFB event to the VNC server. With no argument, simply sends the RFB
equivalent of the X event which caused the action. With arguments, generates
either key or pointer events depending on the arguments:
◊ SendRFBEvent(keydown,keysym)
◊ SendRFBEvent(keyup,keysym)
◊ SendRFBEvent(key,keysym) (short for keydown followed by keyup)
◊ SendRFBEvent(ptr,x,y,buttonMask)
◊ SendRFBEvent(ptr,buttonMask)
where
◊ keysym is the string representing an X keysym. The best way to find
these is to use "xev", or look in /usr/include/X11/keysymdef.h and strip
off the "XK_".
◊ x and y are the position of the pointer event. If not specified, use the
position of the X event which caused the action.
◊ buttonMask is a bit mask representing buttons 1 to 8 with bits 0 to 7
respectively, 0 meaning up, 1 meaning down (pressed). So 0 means no
buttons, 1 means button 1 pressed, 5 means buttons 1 3 pressed, etc.
SelectionToVNC()
Send the local X selection or cut buffer to the VNC server. This is usually
invoked when the mouse enters the viewer window. With no argument or an
argument "new", this is only done if this is a "new" selection, i.e. it hasn't
already been sent. With an argument "always", it is sent each time.
SelectionFromVNC()
Set the local X selection and cut buffer to the current value of the VNC server
"cut text". This is usually invoked when the mouse leaves the viewer window.
With no argument or an argument "new", this is only done if there has been new
"cut text" since the last time it was called. With an argument "always", it is set
each time.
Quit()
Quit the VNC viewer.
Pause()
Pause for a given number of milliseconds (100 by default). This is sometimes
useful to space out events generated by SendRFBEvent.
ToggleFullScreen()
Toggle in and out of full screen mode.
SetFullScreenState()
Sets the "state" resource of a toggle widget to reflect whether we're in full screen
mode.
ServerDialogDone()
PasswordDialogDone()
Used to tell the dialog boxes that entry has finished. Usually invoked by the
return key.
Widget hierarchy
For the real experts amongst you, here is the widget hierarchy used by vncviewer in case
you want to change the resources of the widgets.
Main window:
Vncviewer vncviewer
Form
form
Viewport
viewport
Core
desktop
Server dialog box:
TransientShell
Dialog
Password dialog box:
serverDialog
dialog
TransientShell
Dialog
passwordDialog
dialog
Popup window:
TransientShell
popup
Form
buttonForm
Command/Toggle
button1
.
.
.
.
Command/Toggle
buttonN
WBD White Board Tool
On−line documentation for WBD is available from WBD White Board Tool.
The CD−ROM has a pdf copy of the complete WBD userguide as userguide.pdf.

P2P Application Help

Transcription

Similar documents

Regions CD Check Viewer Installation Guide

Sean Paul - Alex Bienstock

SDSU MASTERS of HOMELAND SECURITY GEOL600 SENSOR

E10 Quick Reference

Network Video Recorder VK-64/VK

SiteManager

STC-CRV2X User`s Manual Toll Free Customer Support 1-877

#104 300 Klahanie Drive • Port Moody

USER`S MANUAL

VNC User Guide