Deliverable (public)
Transcription
Deliverable (public)
Project Number IST-1999-10375 Project Title INTELLECT Document Type Deliverable Document Title Pilot Integration Documentation and Handbook Document Number D13 Due Date 01/02/2002 Delivery Date 01/03/2002 Document Status FINAL Workpackage(s) WP04 Security Public File Name INT-D13FINAL.doc Short Description Keywords This deliverable is a description about the pilot phase within the projects, which is divided into three different user trials. It shows the work of the pilots, the problems which occurred, and provide a handbook for the user and technical understanding. Blauwerk, HiTEC, Interset, Prototype, Pilot, Trial, Implementation, Development Platform Partners owning OPTINET Partners contributed Optinet, BIBA, Blauwerk, HiTEC, Interset Made available to Jorge Gasós (CEC) INTELLECT IST-1999-10375 Letter of Content 1 Executive Summary.......................................................................................................................... 4 2 Introduction ....................................................................................................................................... 5 3 Pilot One: Blauwerk .......................................................................................................................... 7 4 5 6 7 3.1 Requirements of Blauwerk ........................................................................................................ 7 3.2 Surface structure ....................................................................................................................... 8 3.3 Requirements to the developing partners ................................................................................. 9 3.3.1.1 Blauwerk ..................................................................................................................... 9 3.3.1.2 Atlantide ...................................................................................................................... 9 3.3.1.3 BIBA .......................................................................................................................... 10 3.3.1.4 Optinet ...................................................................................................................... 10 3.4 Integration of the modules....................................................................................................... 10 3.5 Feedback of the first prototype version of Blauwerk ............................................................... 12 3.6 Starting of the pilot .................................................................................................................. 14 3.7 Discussion of the missing features with Blauwerk .................................................................. 15 3.8 Set-up of an improved pilot ..................................................................................................... 15 3.9 Conclusion............................................................................................................................... 18 Pilot Two: HiTEC ............................................................................................................................ 20 4.1 Requirements .......................................................................................................................... 20 4.2 Surface structure ..................................................................................................................... 21 4.3 Feedback of the prototype of HiTEC ....................................................................................... 22 4.4 Starting of the pilot .................................................................................................................. 23 4.5 Discussion of the missing features with HiTEC....................................................................... 23 4.6 Conclusion............................................................................................................................... 24 Pilot Three: Interset ........................................................................................................................ 25 5.1 Requirements .......................................................................................................................... 25 5.2 Surface structure ..................................................................................................................... 27 5.3 Starting of the pilot .................................................................................................................. 29 5.4 Conclusion............................................................................................................................... 29 INTELLECT prototype .................................................................................................................... 30 6.1 First prototype.......................................................................................................................... 30 6.2 Current surface structure......................................................................................................... 31 6.3 Second prototype .................................................................................................................... 33 6.4 Common open issues.............................................................................................................. 36 6.5 Conclusion............................................................................................................................... 37 Handbook ....................................................................................................................................... 38 7.1 Architecture ............................................................................................................................. 38 7.2 Installation Guide of the client’s software................................................................................ 40 7.3 Product Data Model................................................................................................................. 41 INT-D13FINAL © The INTELLECT consortium – 2002 Page 2 of 50 INTELLECT 7.4 IST-1999-10375 User Handling.......................................................................................................................... 45 7.4.1 Managing the VR representation of products .................................................................. 45 7.4.2 VR Graphical User Interface for the customer ................................................................. 47 8 Conclusion ...................................................................................................................................... 48 9 Annex.............................................................................................................................................. 49 9.1 Pictures from the whiteboard from Blauwerk’s prototype........................................................ 49 INT-D13FINAL © The INTELLECT consortium – 2002 Page 3 of 50 INTELLECT IST-1999-10375 1 Executive Summary This deliverable is a description about the pilot phase within the projects, which is divided into three different user trials. It shows the work of the pilots, the problems which occurred, and provide a handbook for the user and technical understanding. The prototypes of the project INTELLECT were available since the middle of the year 2001 till the end of the project at March 2002. But there were not all prototypes at the same time available. The first one was Blauwerk in Austria, which has been developed and tested. After that first prototype development, which takes into account new user requirements, the second and third pilot had time problems, because of missing information about the products and adaptation of the 3D objects. Therefore we decided to develop as a second part an INTELLECT prototype which includes all features we want to included. This prototype has been used for the HiTEC pilot from September 2001 till March 2002. After that development the project created a prototype for the Greek company Interset. That prototype has been established from December 2001 till March 2002. Summarised, all pilots run in the last decade of the project successfully, but we had several problems to solve before. This deliverable will describe the different prototypes, the problems, the successful story, and the further user requirements which have been included. Additionally, the technical issues are described. Furthermore, this deliverable will be a guideline and handbook for the used prototypes to give a better impression about the functionality and handling. INT-D13FINAL © The INTELLECT consortium – 2002 Page 4 of 50 INTELLECT IST-1999-10375 2 Introduction The fourth phase of the INTELLECT project brings together the necessary hardware infrastructure including the network connectivity with the developed and selected software system. The information repository has be populated during this task. Simultaneously a scenario environment has been defined to be used for a plentiful testing of the complete system. This testing wants to cover technological and logical functioning of the system. This task has a crucial importance because the pilot has to be matched against its user-friendliness, integration and convergence across information processing, communication and media. The involvement of the end users has been very close. The satisfaction of the end user's needs and the completion of the database needed at the end users sites is a milestone of the pilot phase and the overall project. As a result of the implementation work package the complete set of laboratory tested modules have to be available. As far as they are not already installed at the pilot sites, this has been done as part of this work package. The integration of the overall system has reached in a step-by-step procedure interconnecting the individual subsystems. This includes the integration of the system at the pilot sites with existing database management systems as well as the population of the information repository with a sufficient set of product data. Since the users are consortium members, or members of the user group established during the beginning of the project, they are already familiar with the concept of the INTELLECT system to be implemented. Nevertheless, one of the objectives of this work package is to give them instructions for the usage of the configuration module and to record this in a usual operating instructions booklet. This includes instructions about hardware operation (so far as necessary), software operations and procedures to be followed to achieve the aimed goal. The different pilots can be divided into the following different prototypes: 1. Blauwerk: Austrian pilot, which includes a scooter supplier and seller. The user requirements are to create high quality pages which includes configuration features of the products and multi-media entertainment to create a brand. Neither help-desk features are needed from the user nor direct voice support. A direct order has to manage by the order processing module and should create an e-mail message. The connection to the Internet is only by ISDN dial-up connectivity available. 2. HiTEC: Austrian pilot, which includes a bicycle supplier and seller. The user requirements are to create high quality pages which includes e-shop features and configure possibilities. An help-desk is needed, because of the very different components which are available for the bicycles. A direct order has to manage by the order processing module and should create an e-mail message. The connection to the Internet is only by ISDN dial-up connectivity available. 3. Interset: Greek pilot, which includes a computer supplier and seller. The user requirements are to create an e-shop system which makes it possible to order directly. Also an detailed overview about the ordered components should be integrated. An adaptation of the existing SAP system would be useful by the order system. Help-desk features regarding direct support via voice, video, and mirroring are welcome too. The connection to the Internet is only by modem dial-up connectivity available, but the server can be hosted by another provider. 4. INTELLECT: Common prototype, which includes all modules which have been developed. The surface should be look like the Internet pages of INTELLECT to get a common view. This prototype has been used as a test scenario on the development platform and also as prototype if there will be any problems at the user trials with the respective prototypes available. The following figures present the overview about the work package 4 work plan in detail. INT-D13FINAL © The INTELLECT consortium – 2002 Page 5 of 50 INTELLECT IST-1999-10375 Figure 1: Project overview about work package 4, part 1 Figure 2: Project overview about work package 4, part 2 INT-D13FINAL © The INTELLECT consortium – 2002 Page 6 of 50 INTELLECT IST-1999-10375 3 Pilot One: Blauwerk nd After the 2 review the prototype was not stable enough for using in a user trial. To get a stable prototype we installed a running Linux system at the BIBA’s site and configured all software tools on this single server workstation (also the Oracle database). Next issues were the integration of SSL, NetMeeting as third party tool for voice, video and data sharing, and a perfect adaptation of the frontend/back-end functions. The configurator module from BIBA is working and needed data for a further pilot integration in the first phase. Therefore Blauwerk provided information to BIBA regarding. A better co-ordination between the names/types of the configurator and the user requirements were necessary for the administration and configuration for users/customers. nd The 3D view status was presented at the 2 review and had to be extended by animation and avatar functionality. Unfortunately, Atlantide was not able to handle this, because they are not experience in the creation of 3D objectives and had problems to fulfil the objectives regarding animation and avatar functionality. Therefore, Atlantide had to find a way, either how to create the VRML files, or find an alternative product presentation. The order processing was working and had to be adapted to the user requirements. That included the interaction with the E-Shop module and the user databases. The help-desk system is written for Tomcat 4.X. Therefore we had problems to integrate this functionality into the prototype of Blauwerk. But on the other hand, Blauwerk did not need this feature, because of the nature of their customers. Therefore, the NetMeeting third party tool has been implemented on the INTELLECT pages at http://www.ist-intellect.com and after that in the first pilot of Blauwerk. 3.1 Requirements of Blauwerk Additionally to the user requirements we found out at the start of the project in WP1 there were some more user requirements from the end users within the project INTELLECT as they became an idea what the prototypes could be. Regarding the company Blauwerk the following wishes occurred: 1. Visualisation of single products in 2D must be possible 2. Configurations must be visualised 3. 3D is not necessary for Blauwerk, but would be welcome if it looks good. 4. An avatar is necessary, incl. animations, because Blauwerk wants to transfer a “look & feel” 5. QuickTime or RealPlayer videos or video streaming sessions should be integrated 6. Grey background on each web site 7. Navigation buttons (Elevator buttons) will be placed at the bottom of each subpage 8. The buttons must have an „mouse-over“ effect (change to buttons colour) 9. Quark Xpress is not useful for design adaptation for Optinet; it is more sufficient to use Adope Illustrator to create HTML pages using other prodcuts from Adobe. 10. Configurator button should be on each site; also on the home site 11. Component typs for the city-bike should be available a. Frames b. Wheels c. Fork d. etc. 12. Structure a. Configuration of your personal bike should be available on each site b. Home (Starting Page) the logo should be animated; this must be done by Blauwerk (GIF animation or flash) INT-D13FINAL © The INTELLECT consortium – 2002 Page 7 of 50 INTELLECT IST-1999-10375 c. eShop (order process) – should be merged with the products category (e.g. Willy, City etc.) d. Products: world-of-experience should be shown e. Where to ride 1. Industry 2. City Shopping 3. Freebyke 4. Leisure time / sports ii. Who we are iii. News iv. Configure your personal sidewalker v. Languages (German and English) vi. Sign-up vii. Contact f. Differentiation between a customer and a retailer, B2B login g. Integration of video streams (should be available from Blauwerk): different qualities for different access bandwidths This user requirements had to adapted to the current available prototype of INTELLECT and lead to an open process of development during the whole project. 3.2 Surface structure Regarding the XML/XSL structure of the prototype there are the following wishes from Blauwerk: 1. Products: Overview over the scooter types (9 pieces); for the pilot only one or two types should be configurable, but all types will be shown in an overview (in 2D) a. After that overview a site with details will be shown with a big picture of the chosen scooter b. Buttons will be at the bottom of each page 2. Where to ride: Division in four areas a. Downhill i. Pictures- and video clippings ii. History iii. Animation b. City Shopping i. Pictures- and video clippings ii. History iii. Animation c. Freebyke i. Pictures- and video clippings ii. History iii. Animation d. Camper i. Pictures- and video clippings INT-D13FINAL © The INTELLECT consortium – 2002 Page 8 of 50 INTELLECT IST-1999-10375 ii. History iii. Animation 3. Who we are a. Blauwerk’s story b. Profile 4. News a. Archive b. Newsletter c. Press release 5. Login / Sign-up a. Normal customer b. Retailer 6. Contact a. E-mail b. Site Plan c. Retailer overview d. Telephone e. Opening hours 3.3 Requirements to the developing partners The user requirements lead to the following requirements to the developing partners of INTELLECT and Blauwerk itselfs regarding to fulfil the work in work package WP04. 3.3.1.1 Blauwerk - Adope Illustrator data is necessary at least to set-up HTML pages and adapt this pages to XML/XSL - Approx. 11 page layouts will be created, mainly focused on the products (catalogue) - Product information: i. pictures in 2D ii. Names iii. Short/long description iv. Order Number (internal catalogue number) v. Price, excluding of value added tax 3.3.1.2 Atlantide nd - Much better visualisation of the components as we saw at the 2 - If one component will be replaced, the price has to be adapted, too - Configurable components should be visualised, too - Avatars should be available either in the 3D view or in the animation - Improvement of the administration/configuration - Improvement of the GUI-design is needed INT-D13FINAL © The INTELLECT consortium – 2002 review meeting in Brussels Page 9 of 50 INTELLECT - IST-1999-10375 The currently selected component should be displayed beside the whole product 3.3.1.3 BIBA - Storage of the configuration every time - Better mapping of the product categories with the module - No absolute database addressing, static references in code (arrangement with Anecon) - Playing with the products must be possible (without using the basket); one possibility is to configure own products with an anonymous login - Screenshots from the configurator, 3D module, and eShop for the layout person to get a better impression - Software installation on the user site is necessary to be able to use the 3D applet/view - No static paths, but a use of a configuration file as Anecon implemented - Basket: Infos to the order processing 3.3.1.4 Optinet i. Product group ii. Product number iii. Product name - Standard configuration or new configuration 3.4 Integration of the modules At the user integration meeting in Vienna at July 2001 the first pilot has been started. First of all the different status reports from each module were presented and discussed with the end-users. At this meeting also the detailed project workplan was discussed with the available partners. Here you can read the results of the report from July 2001. The e-shop needs a further work on the back-end system regarding to reach a more stable status. But, the adaptation of the back-end functionality to the front-end site is ready. Additionally an adaptation of the front-end design to a common INTELLECT design has been done. The XML surface divides all visible areas of the surface into objects. That makes possible a very flexible work. Bug-fixes regarding further the adaptation failures are necessary. Also a SSL access via Tomcat and Cocoon has to be integrated for a secure communication between the customer and the e-shop site. Also further implementing of more back-end functionality regarding the user requirements is wanted. At the first step the HTML sites had to created, because the end-user Blauwerk was not able to create this pages on their own without help. Therefore a complete concept was developed from Optinet for the end-user’s web-sites. Afterwards the adaptation of the created HTML pages to XML/XSP had to be done. The virtual reality module developed no progress since the last integration meeting in May. That means in July there are no development available regarding the avatar or the animation of bicycles or scooters. Also, not any adaptation has been done of the other modules to the existing user requirements. Atlantide promised to establish the development of the avatar and the animation as soon as possible. Additionally the 3D data will provide by Atlantide if other ways are not possible (e.g. to get this data from the producer or the user). The implementation of the Java 3D applet into the web site and 2D view should be possible too. But this features are still missing till today. The configurator consists of the new product data in the database. Therefore the city roller is complete, but another scooter is still missing. No problems occurred during the implementation. But more information were not available. The information inside the database can be used for the e-shop INT-D13FINAL © The INTELLECT consortium – 2002 Page 10 of 50 INTELLECT IST-1999-10375 module too. Furthermore a configuration file is necessary for all static data (e.g. database connection, type name, etc.). The help-desk system adapted the third party tool NetMeeting for the INTELLECT pages under http://www.ist-intellect.com. We used one official IP address from BIBA to test the implementation carefully. In July open questions were the direct agent support and the adaptation of the mirroring mechanism. No further progress regarding the mirroring was available, because of no progress in Tomcat 4.x and Jboss integration. The order processing has been extended of the basis functionality. An e-mail can now include a detailed list of components for every ordered product and further content of the order. Arbitrary configureable sequence status is also available. Branch within the workflow with alternative sequence status is possible. Also error handling and an auto-Trigger event (for each status can be defined as an event) have been foreseen. But, the working together with the database of Itchy (the development platform) seems to be a bit buggy. Also there is missing content (text) from Blauwerk regarding the email messages. The web-site has been extended with new articles, information, and a help-desk function. This means, the help-desk third party tool NetMeeting is available on the INTELLECT page under the question mark of each site. This functionality is only for test possibilities available, before we integrate it into the INTELLECT system. Also the prototype is available by a direct link on the site, which will show the INTELLECT system with its own design. Figure 3: Public site of INTELLECT with updated information The problems with NetMeeting regarding NAT, firewalling, and dynamical port numbers are mentioned on the help-desk web-site (see Figure 4). Also the tool NetMeeting is available there. The web-sites of the INTELLECT project has been also published on a CD-ROM at September. Here, people who are interested in can get information of the project (in a similar way as they can get information by the web-sites). Additional information will be a photo gallery of each partner meetings and some small videos. Each partner gets one or more CD-ROM(s) for marketing activities continuously. INT-D13FINAL © The INTELLECT consortium – 2002 Page 11 of 50 INTELLECT IST-1999-10375 Figure 4: Help-desk site with integrated NetMeeting The outcome of that first pilot phase was that further adaptation of the modules is necessary to get a stable prototype as demonstration platform. The set-up of the next pilots will be the same work as at Blauwerk. The partners have to collect data and information after the same example. 3.5 Feedback of the first prototype version of Blauwerk After the development of the prototype for Blauwerk the first feedback of them were necessary to improve our development and integration further. The Figure 5 and Figure 6 shows the results of the first realisation. Figure 5 shows the start page which is based on a “shop-xml” file as you can see. From here you can go inside the different pages, which have been defined at the first workshop at Blauwerk. INT-D13FINAL © The INTELLECT consortium – 2002 Page 12 of 50 INTELLECT IST-1999-10375 Figure 5: Home site from Blauwerk Figure 6: Where to ride site from Blauwerk INT-D13FINAL © The INTELLECT consortium – 2002 Page 13 of 50 INTELLECT IST-1999-10375 The remarks from Blauwerk were (more or less) available regarding design and feeling. This end-user is very strong about things like subjective feeling and handling of its own products. As you can see the current pages under http://www.blauwerk.at there is a big progress available to the new designed pages. The following notes came from Blauwerk regarding their new surface: - Arial should be used as font type - The same font type size for all pages - New configure logo should be send as new picture from Blauwerk - “Sign up” will be “login” with two submenu “login” and “sign-up” - New navigation button image should be send from Blauwerk - Every button has the same size - Home button with Blauwerk logo will be added to every site as the first button of the navigation list - German version should be developed in the future - “AGB” submenu should be added to “shop now” - “Where to ride” should be used instead of “See it,feel it,do it” in every site - The name of the active site in the navigation list should be showed in red - The button of the active site in the navigation list should flickering like a heart, when this is too difficult, then use the button with grey middle point - Large photos should be send from Blauwerk - The large photos should be showed in a new window with a size over 250x150 - High resolution and low resolution of the video should be showed in a new window with the size over 200x150 - “See yourself ride” site will established similar as the configuration site - New design for the site „community“, “basket“, “order“, “check out“, “thanks for shopping“, and „no shopping“ should be sent from Blauwerk - Sub-menu “production info” should be added to “who we are”, and the design should be sent from Blauwerk - In the site “cityLTD”, the link with the site ”citySFD” should be added by the cityLTD image - Each text for the site “contact”, ”news”, and “who we are” should be sent from Blauwerk 3.6 Starting of the pilot After the user integration meeting the pilot started at the end of July till now. Only the client site has to be prepared for the 3D view, because we handle all other things separately on the server site (development server Itchy at the BIBA institute). The installation on the client site is not very useful and a little bit complicate. Therefore, Optinet supported the user by the installation process continuously. During the installation of the client software on a Microsoft Windows ME system occurred the following problems: o Problems with the Java 3D extension o No OpenGL 1.1 support at Microsoft Windows ME (Windows NT with service pack 3 includes OpenGL 1.1) o Java 3D is not able to install on Windows ME, maybe because of missing rights on the Java Environment We had to do some more tests for Windows ME, but it was necessary that Blauwerk switch to Windows 2000, because of the existing problems with Windows ME. In July we had still no 3D data INT-D13FINAL © The INTELLECT consortium – 2002 Page 14 of 50 INTELLECT IST-1999-10375 which fits to the database data. Therefore, it was not able to start the pilot with real data! This is also still the problem at the end of the project. 3.7 Discussion of the missing features with Blauwerk Regarding the different modules Blauwerk discussed several improvements with the developers directly. Especially, the order processing and the virtual reality are from special interest for Blauwerk and had to be discussed: 1. Order processing a. The whole order takes one way from customer to the supplier; the order is orderbased not product-based b. Splitting from configured products and standard products should be possible, because the standard products are available on special sites worldwide and the configured products should be handled in Vienna directly c. E-mail orders can not be handled if the order increase: this must be automated, e.g. with a ERP system like SAP 2. Virtual Reality a. Animation of the avatar is necessary for Blauwerk b. A few parameters should be available i. Size of the avatar ii. Male and female avatar c. 3D visualisation is not important for Blauwerk; therefore a 2D visualisation must be able to use d. Is it possible to integrate the Java applet into the web page? i. 3D data are not available by Blauwerk and probably also not by the other user partners. ii. Therefore, Atlantide has to provide “dummy” 3D data iii. Copyrights will not be a problem, because we can use this data only for the prototype (well enough for marketing or advertisement from the provider point-of-view) Another possibility was discussed with Atlantide regarding the extract of 3D data from 2D photographs. Some tools exists and have been tested. But there was no enough user support from Atlantide to get this data in a well enough quality from 2D data. 3.8 Set-up of an improved pilot After the missing parts were integrated (but something is still missing), the next prototype for Blauwerk was established. Find here the new style and surface of the improved prototype. Also the surface was new structured and several templates were foreseen to get better possibilities to adapt the next pilots in less time than the first one. The technical description you can find in the handbook in this deliverable. INT-D13FINAL © The INTELLECT consortium – 2002 Page 15 of 50 INTELLECT IST-1999-10375 Figure 7: Start page of Blauwerk Figure 8: Login mask for the customers INT-D13FINAL © The INTELLECT consortium – 2002 Page 16 of 50 INTELLECT IST-1999-10375 Figure 9: Template for getting the user profile Figure 10: Product which can configured and add to the basket INT-D13FINAL © The INTELLECT consortium – 2002 Page 17 of 50 INTELLECT IST-1999-10375 Figure 11: Client VR configuration applet Configuration of the product is accomplished as follows (Figure 11). The user can select the component he wants to modify in the components list or in the viewing area, if he is in the interaction mode, and if the 3D model is available. The customer indicates then, that he wants to replace this component, thereby opening the component exchange dialog (see Figure 12). Figure 12: Component exchange dialog The component exchange dialog displays the list of possible components that can replace the selected one. This component list is generated by the Configurator from the product database and represents the subset of available components that will fit into the product physically and mechanically and at the same time meet all the “soft” criteria specified for this product, or manually specified by the user. Selecting a component in this list shows a 3D view of the component (if available) and its technical description (see Figure 12). 3.9 Conclusion The meetings in Vienna were necessary and useful to go a big step further to complete the first prototypes and to start the prototypes. Nevertheless after the first sessions the developing partners got their work to establish the prototype without bugs and limits. Furthermore, the experience from the first pilot was very helpful for establishing the next pilots at HiTEC, Austria and Interset, Greece. The project has saved time for this next project phases. Get at the next two figures some impressions of the meetings in Vienna. INT-D13FINAL © The INTELLECT consortium – 2002 Page 18 of 50 INTELLECT IST-1999-10375 Figure 13: Sitting together at Blauwerk’s office for the client preparation Figure 14: Scooter from Blauwerk INT-D13FINAL © The INTELLECT consortium – 2002 Page 19 of 50 INTELLECT IST-1999-10375 4 Pilot Two: HiTEC After the first pilot has been started the next pilot in Austria has been established. Because of the missing 3D objects, and the new developed web pages of the end-users, in the first phase we used the INTELLECT prototype for the user trials. There was a widely underestimated problem of getting/generating 3D data for the product components to be configured: 1. The first reason for this was, that 3D models would have had to be based on real construction data, the details of which the suppliers did not want to make public. 2. Secondly, there were no affordable and easy-to-use tools to generate 3D data from the component data, which the end-users had available. 3. Thirdly, any substantial outsourcing of rendering tasks to professional suppliers would have resulted in massive overspending of the project budget. 4. The attempts to collect appropriate 3D data from suppliers required much of the resources initially planned for testing. Numerous contacts to the top-level managers of international companies like shimano or manitou had to be established and many discussions were held but yielded no results. 5. There were also attempts or to generate 3D data within the consortium. They were based on existing plans and additional 2d constructions made specially for the project. These data, however, also turned out to be no sufficient basis to create the necessary 3D models. In that cases the most of the Testing had been done on a generic level with the INTELLECT prototype shop with generic 3D data objects. The products of the end-users were de-composed for the configurator and all the rules for combinations have been defined. The end-users like HiTEC redesigned and re-launched their web-sites to provide the foundations for the shop-to-be. 4.1 Requirements The requirements from HiTEC are very similar to the requirements of Blauwerk. They want to improve the 3D visualisation like Blauwerk, too. But against Blauwerk, they know how to create HTML pages and use new web technologies like Flash. Because of a complete re-design of their web-pages, which has been done by HiTEC on their own advise, the project INTELLECT had to wait on this results. That means we had additional delays for this pilot in Austria. The following requirements from the project to HiTEC are: - HTML pages for the pilot are necessary - Information about the products is necessary for the database - Information about the rule structure for the configurator - 2D photography from the products for the 3D modelling - Data for workflow and email text is needed for order processing And there were also special user requirements from HiTEC to the development partners: - Better visualisation of the 3D view, because of they high-end bicycles - Sufficient handling of the configurator components - Fast access to the Java3D window - High-end pictures of the bicycle’s components - Help-desk system for a direct support during the choice/configuration of the product - Basket functionality and overview about the chosen products The big advantages of the end-user HiTEC was that they develop their own pages without direct help of the developer partners. Therefore all other partners had time to develop the back-end in a more perfect way. INT-D13FINAL © The INTELLECT consortium – 2002 Page 20 of 50 INTELLECT IST-1999-10375 4.2 Surface structure As you can see in this surface structure there has been foreseen INTELLECT parts like the configurator, basket, and user login. All pages are available as Flash and HTML. Figure 15: Start page of the HiTEC web pages Figure 16: Configuration page of HiTEC INT-D13FINAL © The INTELLECT consortium – 2002 Page 21 of 50 INTELLECT IST-1999-10375 Figure 17: Product in detail Figure 18: Contact page for user profiling 4.3 Feedback of the prototype of HiTEC Because of the self development of the web pages there was no feedback necessary. For the product data and the integration of INTELLECT features there were an exchange of information continuously. INT-D13FINAL © The INTELLECT consortium – 2002 Page 22 of 50 INTELLECT IST-1999-10375 4.4 Starting of the pilot The pilot has been started in September 2001 with the prototype of INTELLECT. Therefore, only a scooter and different PC components have been used. For the technical description please read the chapter about the INTELLECT prototype. So, the experience of the pilot was, as here mentioned, based on a generic level. The mountain bikes produced by HT Trading consist of parts purchased from renowned suppliers like Shimano, Marzocchi, Manitou. When the representatives of these producers were contacted for the first time the first half of the year 2000 with information about the Intellect Project, there were positive and encouraging reactions. The industry appreciated the potential of promoting high-end equipment through appealing VR/3D representation. Industry representatives especially stressed the necessity of a detailed presentation of the product features to justify the higher prices of professional components. When the first prototype had to be filled with real data, the first problems were encountered. According to the producers there were component data available, both in 2d and 3d format. Some of the companies even had rendered product files for simulation purposes. However, the company representatives were reluctant to give this information to outside parties. HiTEC contacted suitable producers on numerous occasions: - Marzocchi Spa (Bologna/IT) was contacted on the Taipei Cycle Show in April 2002 on management level. The result was, that the company would reconsider their point of view – after a few weeks an numerous meetings with the Austrian representative, HiTechs request was turned down, - A formal meeting with Shimano Europe Gmbh. (the European branch of the world market leader in gear shifting and brake equipment) took place at the “EuroBike” Fair in Friedrichshafen. Also the European General Distributor participated in this meeting. Again no decision was taken at this event, but was postponed for further internal consultations. After a few weeks and two meetings with the head of the marketing department Mr. Zeilberger only a CD with component images was sent with an apology that this is all what Japan was ready to distribute. - For the negotiation with Answer Products USA (Manitou etc.) HT was in a better position, being the general representative of the company in Austria. The first request for data, with which 3d models could be generated, was communicated to Answer at the Taipei fair of 2001. Through the regular contact with the European Management the headquarters in the States was pushed even more. Nevertheless the request was rejected. Only after an intervention in a personal meeting with Glen Miller, Answer’s president an assent could achieved. It took until 11/2001 until one file with the construction data of a 2 two year old fork reached HiTech. The situation is maybe bets explained by a mail sent to HT by Manitou Inc. during the negotiation phase, in which the development department explains that “.... any competitor would be able to use the data for copying our product to the very detail. It even would be able to feed the information to production machines and get a bootleg product within a few hours”. 4.5 Discussion of the missing features with HiTEC The lacking co-operation from the suppliers was a major drawback for the overall goals of HT and endangered the whole strategy of bike configuration via the Internet. As a consequence HT Trading tried to get 3D data from other sources. First there were attempts to capture construction data by copying them by a CAD system. HT technicians, though experienced with system used hours of drawing time without promising results. 2D construction plans could be generated quite easily, because it was similar to the development work being done for the components designed by HT personnel. When it came to adding the third dimension to these drawings, the problems could not be overcome. Firstly an abundance of additional data would have been necessary to give in. Secondly, additional program components would have been necessary together with lengthy qualification measures for the employees involved. rd All in all, no economic way of generating 3D in-house could be found. So, contacts with some 3 party suppliers of rendering services were established. But very quickly it became apparent, that outsourcing the production of 3D models of existing parts would also become a very expensive task. INT-D13FINAL © The INTELLECT consortium – 2002 Page 23 of 50 INTELLECT IST-1999-10375 There were bids for creating a detailed 3D model of key parts like gear shifts components at about 3.000 Euro, which would rise the price of a fully equipped VR-mountain bike summed to over 20.000 euro. As a potential alternative, there were contacts to a company providing 3D scanning services. But also in this case, the cost was way too high for a economic production of the necessary data (app. 5.000 Euro per hour including personnel, with the potential outcome of 1-2 parts per hour). Also within the consortium there were attempts to generate the necessary data for the visualisation module using. Existing blueprints were transferred to CAD systems and tested by development partners for their usability as source material for rendering. In anticipation of the configuration capabilities, the web-pages of HiTEC was completely redesigned. The company’s strategy to offer different variants of the same bike was already apparent in the recent versions of the site, but different components were only shown in text form. The new web-pages was designed for highlighting the different variations and to show a picture (maybe later a 3d model) of every key component. There was the goal to provide a clear structure for the configuration of different bikes based on standard models, which could be upgraded/stripped down by the prospective buyers. This aspect was also considered by the organisation of the bike’s components, which was elaborated for as a basis for configuration module of the Intellect Shop. It had to enhanced from a simple part list to a multi-level component structure, where each level consists of – interchangeable – parts, which are combined to modules being the parts for the next level. The conception and development of the overall web-pages was done in in co-operation and with support of the developing partners, but also yielded problems. HT’s long-standing local web designer had foreseen a different technical background (flash), which was hard to fit into the technical concept if INTELLECT, which is based on XML. That caused a delay for the establishment of the pilot with the original pages and is not solved yet, because of the less time till the end of the project. Because of the incomplete prototype of HiTEC, there are some important features still missing: - Adaptation of the HTML pages to XML/XSL pages - Integration of the SSL - Integration of real VRML data (3D objects) - Integration of help-desk functions (NetMeeting and mirroring) - Further extension of the database content Other features like the configurator, user login/profile, and 2D visualisation have been complete integrated. 4.6 Conclusion Summarising it might be said, that there was a massive under-estimation of the efforts, which would be required for getting the basic data for a VR/3D shop in general. The suppliers had understandable objections against giving away construction details, which could be used for copying the components, But without their co-operation, a different concept would have to be used. Maybe economic 3D scanners would have solved the problem. Or, if he 3D representation aspect would have been foreseen in the construction process, more simple and not so secret-revealing CAD files could have been used. The consortium also had to learn, that maybe 3D technology available at the moment is still too sophisticated and costly to be used in a plain commercial project. There was no way to justify the additional costs for creating VR promotion material, since it would not necessarily yield so much additional turnover/profit to break even within a reasonable time frame. 3D technology - when not already foreseen in the construction process and therefore produced only for marketing reasons – would almost double product development cost and would therefore considerably reduce the producer’s margin. INT-D13FINAL © The INTELLECT consortium – 2002 Page 24 of 50 INTELLECT IST-1999-10375 5 Pilot Three: Interset The prototype for the pilot of Interset has been finalised with a delay too. That reason was that also here the re-launch of the web-pages needed time. Interset developed and designed there web pages on their own. Therefore no navigation structure, additional features/requirements, or design has to be developed from the partners of INTELLECT. Under advise of the partners they established their webpages with the additional features of the project. 5.1 Requirements The user requirements from Interset did not changed since the first definition in WP01 has been started. Therefore nothing had to new adapted. Additional they don’t need 3D visualisation, because of their products (it is not necessary to show a motherboard or a graphical board in 3D). The requirements from the project to Interset were: - Information about the web-sites (if possible as HTML) for the pilot - Information about the products - Information about the rule structure for the configurator - 2D photography from the products for the 3D modelling - Data for workflow is needed for order processing The technical description of the products is very deep. The next table shows a Ultra SCSI connectivity, BIOS on card, data transfer rate up to 20Mbps. This input is available in the database to make a configuration possible. Why to buy Boost your CD-recordable or Jaz drive to its best performance System Environment Desktop PCs Key Differentiators Ultra BIOS EZ-SCSI Deluxe Edition SCSI on connectivity card Data Transfer Rate Up to 20MB/sec External Connectors 50-pin High-Density (SCSI-2 female) Internal Connectors 50-pin standard (male) Bus Type 32-bit PCI System Requirements INT-D13FINAL © The INTELLECT consortium – 2002 Page 25 of 50 INTELLECT IST-1999-10375 IBM-compatible PC-486, Pentium or above Windows 2000, ME, NT, 95/98, 3.1, or DOS operating systems PCI expansion slot SCSI peripheral Package Contents SCSI EZ-SCSI 50-pin internal Quick Reference Registration card Card SCSI cable (external installation cable not 2930U software included) guide guide Warranty 5-year manufacturing and materials warranty Picture: Another technical specification of a motherboard are: ® TM Supports Intel Pentium III/II CuMine, Katmai and Celeron processors, 2 x DIMM for up to 1GB PC133/HSDRAM (high speed), UltraDMA/66, 5 x PCI, 1xAGP. The technical specification is: ® The ASUS CUV4X-C Mainboard is based on the VIA 694X chipset with ATX form factor for the latest ® TM support in Intel Pentium II/III 300~1GHz+ Coppermine and Celeron processors. This fascinating chipset is equipped with 133MHz Front Side Bus (FSB), and support for PC133 SDRAM and AGP 4X, Ultra DMA/66, Wake-On LAN and Ring. It is also bundled with ASUS PC Health Monitoring to monitor and ensure maximum safety for your PC. Chipset Feature ® VIA 694X is a high performance chipset which uses 82C686A South Bridge to offer 133MHz memory bandwidth for supporting wide range of PC133 memory devices. • Supports Intel Pentium III/II CuMine, Katmai and Celeron • 133MHz Front Side Bus • 5 x PCI, 1 x AGP • 2 x DIMM for up to 1GB PC133/HSDRAM (high speed) • UltraDMA/66 • PC Health Monitoring ® INT-D13FINAL TM © The INTELLECT consortium – 2002 processors Page 26 of 50 INTELLECT IST-1999-10375 Picture: That information shows that 3D is not useable in this pilot and there are many database information available for all the different components and categories. 5.2 Surface structure The surface structure of the pages of Interset were set-up by Interset! They also adapt their HTML pages to XML/XSL code. Optinet had to only advise the partner regarding the additional features of the INTELLECT e-shop system and had to adapt the XML structure of Interset with the structure of INTELLECT. Figure 19: Start page of Interset INT-D13FINAL © The INTELLECT consortium – 2002 Page 27 of 50 INTELLECT IST-1999-10375 Figure 20: Interset’s news page Figure 21: Product overview sorted for manufactures INT-D13FINAL © The INTELLECT consortium – 2002 Page 28 of 50 INTELLECT IST-1999-10375 Figure 22: Product list in the basket 5.3 Starting of the pilot The pilot has been started at January only. The reasons are that the re-design of the web-pages has been done by Interset and it caused much time. The last input came in December 2001 from Interset. After that input Optinet had to adapt the XML structure of Interset with the common structure of INTELLECT. Also the additional features like the help-desk system, basket functionality, shopping cart etc. had to adapted and needed time. Before the Interset pilot started they collected experience with the common prototype of INTELLECT. Another problem was the 2D configuration which should be possible for this prototype. This mean a complete adaptation of the configurator and the 3D visualisation. Therefore, additional time was needed. Nevertheless, the pilot is working since January 2002. There are still some open points which are the same as in the other pilots. 5.4 Conclusion The workshop in Greece is planned for the end of February, based on the results of the outcome of the existing pilots. This workshop is an important step (also for the Commission) to show that we are using the chance to involve potential customers. Target companies are customers of Interset (or similar users like Interset) or potential users of other areas (kitchen, furniture, garden, insurance business). In addition some strategic partners to establish the value chain will be invited. Interset want to use the prototype as a real product for their own web pages and they want to offer their customer the features which they wanted. INT-D13FINAL © The INTELLECT consortium – 2002 Page 29 of 50 INTELLECT IST-1999-10375 6 INTELLECT prototype There were different INTELLECT prototypes available during the project. The first of all you can see at nd the Figure 23 and Figure 24, which has been presented in Brussels at the 2 review at May 2001. The current surface you can find at the chapter afterwards. Within the prototype you can find the whole functionality of the prototype which can be adapted to each pilot of the project on a simple way. Since the end of 2001 also the new version of the software environment is available: Tomcat 4.0. The use of these tools makes it necessary to implement a new application server. In addition this would be needed for the integration of the mirroring with the whole shop. All software is licence are free (if you replace Oracle by Postgres SQL for example). 6.1 First prototype nd The first prototype which has been shown in Brussels during the 2 review of the project was not a stable software platform, because the different modules run on different machines and were not really integrated in one platform. Also the surface was not very useful or got not a well design as you can see in Figure 23 and Figure 24. Figure 23: First start page which was presented in Brussels during the 2nd review The next page is the catalogue of the products you can select for configuration. In the first prototype we integrate only one product in three different categories. That means, you can select a computer as a “slow”, “medium”, or “fast” one. After that choice it is possible to configure the product by the fact that you can select different components which fits into the whole product. In this phase additional features like real 3D view, shopping cart, user profiling, and different language support were not integrated. But it was a big success that the different developed modules fit into one platform and play with each other without big trouble. INT-D13FINAL © The INTELLECT consortium – 2002 Page 30 of 50 INTELLECT IST-1999-10375 Figure 24: Product catalogue page from the first prototype 6.2 Current surface structure The project INTELLECT has the requirement that we want to support different users (e.g. ISPs, ASPs), who wants to offer their products to their customers. Therefore, we need a flexible user interface which can be adapt on each user requirements which can be also changed during a project’s lifecycle. Therefore, for the surface of the prototypes the e-shop pages are divided into several areas. The requirements for a highly customisable localisable user interface are: - All content must be translatable and customisable (the static textual content of UI and the dynamic content coming from the dB). - All names (action names, attribute names,....) and their explanation (=description) must be translatable and customisable, (e.g. the attribute "date" appears on a German page as title of a column like "Datum") - The translation or customisation of any name, any explanation of a part of that object (attribute,....) or it’s icon can be made in a central place just once and must be reflected immediately and consistently in all the places where this name is shown - Presentation mechanisms (there are just a few, e.g. Tree View, List View, Detail View, Presentation of Actions (Pull-Down Menus, Pop Up Menus), Choice of alternatives....) must be used consistently in the whole USI. The Look of each presentation mechanism shall be customisable in one central place (e.g. bullets or “+” symbol for tree view) - Consistent appearance of an object, wherever I see it. Restrict what I see from an object for a specific appearance of that object in a USI - The general look can be changed / customised in a central place just once and must be reflected immediately and consistently in all the places (e.g. Colour, Font, ...) - It should also be definable, which presentation mechanism is used if there are alternatives (e.g. possible actions presented as pop up menus or pull down menus; choose of alternatives via combobox or radiobuttons) - Support of and specific customisations for different client technologies (WML, HTML,...) INT-D13FINAL © The INTELLECT consortium – 2002 Page 31 of 50 INTELLECT IST-1999-10375 Navigation: NEWSFLASH: MainMenu MainMenu Headline New Bike Race in Vienna Special Brakes available Bla Bla Bla Tralalalalala MainMenu MainMenu SubMenu SubMenu SubMenu MainMenu MainMenu LOGO 2 Date 1.7.00 2.6.00 1.6.00 9.5.00 Click on the News article to see details below! LOGO 1 SubMenu Location Vienna Oslo Rennes Bremen Selected News article: New Bike Race in Vienna Picture 2 Vienna, 1.7.00: Bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla bla Order Tracking You already ordered products from us and want to know, if it has already been delivered? Log in here: User: Password: Special offer of this week: Get the new Shimano Break for only $ 245,-!!!! Click here to get details.... Picture 1 Forgot your password? Click here! Figure 25: Surface areas for XML/XSL support The customisation of the e-shop has two different aspects: one is the overall appearance of the shop which has to be defined once before the shop is on line and the second is part of the contents which can be variable. The difference between these two is the method. The appearance for the e-shop can be defined in a configuration file, the content can be changed with the administration tool. The following parameters can be changed: 6. for the appearance: a. Colour b. Font c. Logo d. Languages and currencies, how and where to switch e. Sound f. Layout: where are the navigation toolbars, the frames for the main information etc. g. Navigation possibilities: links to other sites, menu items h. Appearance of menus: pull down, pop up, choice of alternatives with radio buttons or combo boxes i. Use of flash / or not j. appearance of the catalogue structure k. Search possibilities, e.g. for manufacturers, categories or products, only with headword l. Payment methods m. General Trading Conditions n. Delivering methods 7. for the content: INT-D13FINAL © The INTELLECT consortium – 2002 Page 32 of 50 INTELLECT IST-1999-10375 a. Fields for the customer data b. Catalogue structure, e.g. new categories c. Attributes for new product categories d. Changes of the navigation, e.g. new menu items e. The configuration of the shop should be as flexible as possible but if the shop operator prefers a static configuration, it might be a good idea to configure all items with the configuration file. Pages like in Figure 25 can be realised in many ways, the classical approach to achieve this would be to code every page in the shop manually, that means the HTML with it’s included dynamic tags would be individually edited. If there is another merchant, the pages can be copied and changed to the other merchant’s needs. The back-end functionality will not be changed – they will only be used or not. This approach includes the following two major ideas: • Using XML (eXtended Markup Language), XSL (eXtended Stylesheet Language) and DTD (Document Type Definition) as basis to support the requirements of different merchants and different client technologies. • Object centred GUI definition, customisation and implementation • Define USI representation not for a page or a frame, but define them per object, which has to be represented and can be reused wherever the object appears in which representation ever. • Choose the objects you want to see in a frame / on a page. Place them. Choose the presentation mechanism (list view, detail view, combination of list and detail view,...). Restrict the default appearance if necessary • The View of the object can be different for different user roles (e.g. the view of an object is different, if an administrator, help desk agent or a normal customer gets the page) The page creator can use an XML-based page description language, where he defines the parting of the page into different areas (frames, boxes,....) – this step is similar to defining the frameset for a multiframe web-page – and then define an object and it’s presentation mechanism (e.g. Product Details or Product List), that is responsible for producing the content for this area. 6.3 Second prototype The Figure 26 shows the end result of the surface of the INTELLECT prototype. This is an common surface which will be used for common advertisements or project presentations. All features are available together, while that is not the case at the pilot prototypes, because of the different requirements. The configuration is only possible with abstract 3D objects as you can see in Figure 27. The session ID in Figure 26 helped us to recognise if the same user is shopping or changed during the session. The features which can be used are: - Configuration of three products of the database - Login - User profiling by a login template and the same session ID - Help-desk functionality by the integration of the third party tool NetMeeting - Basket - Shopping Card - Catalogue - Multi-language support - Order processing - Visualisation of abstract 3D objects - XML/XSL surface for fast adaptation of new requirements INT-D13FINAL © The INTELLECT consortium – 2002 Page 33 of 50 INTELLECT IST-1999-10375 Figure 26: Surface of the INTELLECT system Figure 27: Virtual Reality Java Applet Window Figure 28 shows the catalogue page which includes additional one scooter. All information you see come from the database as scooter description, producer logo, and price. There is no static information more available against the first prototype. On the right hand you see the help-desk feature which includes a videoconference windows if there is an agent available on the other site. Here you can see only the NetMeeting logo, but it is direct integrated into the web page. INT-D13FINAL © The INTELLECT consortium – 2002 Page 34 of 50 INTELLECT IST-1999-10375 Figure 28: Product catalogue of the INTELLECT prototype The last figure of the current prototype shows the user profile template, which must be used to get a login/password information by the provider of the e-shop system. This template can be adapt to different user on a simple way. Also here you can get help directly or look at your basket. INT-D13FINAL © The INTELLECT consortium – 2002 Page 35 of 50 INTELLECT IST-1999-10375 Figure 29: Entry template for the user profile 6.4 Common open issues Regarding the development and integration there are still open points which will be solved after the project time if the prototype of INTELLECT will be a real product. Therefore we can divided the needs of the future into the different modules of INTELLECT: 1. eShop (Optinet) a. Producer info (by the users) b. My configuration (add to basket, delete configuration, configure configuration) INT-D13FINAL © The INTELLECT consortium – 2002 Page 36 of 50 INTELLECT IST-1999-10375 c. Pictures and content (from the user of INTELLECT) d. Community building e. Merchant support (is foreseen) f. Developing an administration tool for the e-shop products 2. Virtual Reality (Atlantide) a. Get 3D data from the users or if that is not possible integrate “dummy” data (very important for the pilots and the next review) i. VRML.JAR file is necessary in the first step for the current pilots for two products; the current version is not sufficient for the pilots! ii. 2D photographs from the users b. Avatar integration/extension c. Extend the 3D visualisation with animation of the Avatar d. 2D visualisation of the products (bigger effort as the 3D view, because of the more pictures which are needed) e. Integration of the 3D applet within the web-site f. Better and more efficient client adaptation 3. Help-desk (Anecon) a. Integration of the mirroring mechanism in Tomcat 4.x b. Integration of the administration tool c. Efficient handling of incoming calls 4. Order processing (Anecon) a. Handling of orders for different suppliers 6.5 Conclusion The common prototype of INTELLECT is stable enough to work in different environments for different users with different requirements. Therefore, the project members want to set-up a real product after the project’s lifecycle. To get a real product, there are still some open points available which have to be developed if such a product will be successful on the market. A real problem is the 3D visualisation from the user and the provider point-of-view, because it is not possible to get 3D pictures direct from the producer as we got the experience in our project. Also it is not possible to get this information by the provider (e.g. the format of the 3D data is not the same, there are no 3D data available). Therefore the offering company has to create this 3D data on their own. That is in many cases time critical. On the other hand, the end-user clients must be prepared for the 3D view, because Java3D is not integrated in today’s browsers. That is not very comfortable at the moment. Hopefully new browser version will support this technology, otherwise it will not be successful. Finally the software versions were not stable enough in the past to create a real product. That must be changed in the near future. The handling with the application server is also not fast enough. Bigger bandwidth and faster server platform should be necessary for the provider, but also for the end-users. The results you can find in detail in deliverable D15. INT-D13FINAL © The INTELLECT consortium – 2002 Page 37 of 50 INTELLECT IST-1999-10375 7 Handbook This description will give an overview about the complete architecture and the handling of the different modules. 7.1 Architecture The INTELLECT consortium provides an enhanced eShop system including all practicable electronic commerce features, and offering a suitable and realistic representation of the products. The INTELLECT application contains five modules: 1. The eShop module takes care of the general functionality of an e-commerce system (basket, catalogue, secure communications, localisation, multimedia presentations, …). 2. The Virtual Reality module provides 3D visualisation of the product, and enables the customer to configure its own product by choosing each of the components. 3. The Configuration module manages the configuration possibilities. 4. The Help-desk provides on-line help to the customer, using features like chat, voice over IP, and videoconferencing. 5. The Order-processing module manages the orders, and implements the interface between the INTELLECT application, and the back-office systems. The INTELLECT product is based on a three tier (layer) design. The next figure provides an overview of the software products and tools, which were necessary for the development of the prototype of INTELLECT. On the one hand, the prototype uses open source products as it's underlying backbone. On the other hand, third party products are used for direct real-time communication via NetMeeting or security connectivity via SSL. In addition, commercial products regarding databases are used. Therefore the architecture can be devided into three tiers on the server site: 1. JBoss (Tier 1): The JBoss/server is an open source, standards-compliant, Enterprise Java Beans (EJB) application server implemented in 100% pure Java code. The JBoss community of over 500 developers world-wide is working to deliver the full range of J2EEtools as the premier enterprise Java application server for the Java 2 enterprise edition platform. The JBoss/server and complement of products are delivered under a public license. With 1500 downloads per day on average, JBoss/server is the fastest growing J2EE based server worldwide. 2. Tomcat (Tier 2): Tomcat is the reference implementation for the Java Servlet 2.2 and Java Server Pages (JSP) 1.1 technologies. It includes the web server technology and offer this functionality base for the prototype. 3. Cocoon (Tier 2): Cocoon is a 100% pure Java publishing framework that relies on new W3C technologies (such as DOM, XML, and XSL) to provide web content. The Cocoon project aims to change the way web information is created, rendered and served. The new Cocoon paradigm is based on the fact that document content, style and logic are often created by different individuals or working groups. Cocoon aims for a complete separation of the three layers, allowing the three layers to be independently designed, created and managed, reducing management overhead, increasing work reuse and reducing time to market. 4. Oracle (Tier 3): The database Oracle8i standard edition enables business to cost-effectively develop and deploy Internet-based solutions. Oracle8i includes native support for technologies like SQL, XML, and Java that have become the standards on which today's e-businesses are built. For the independence of the end-user presentation and functionality, the project decided to use Java Server Pages (JSP) technology which allows web developers and designers to rapidly develop and easily maintain, information-rich, dynamic web pages that leverage existing business systems. INT-D13FINAL © The INTELLECT consortium – 2002 Page 38 of 50 INTELLECT IST-1999-10375 As part of the Java family, JSP technology enables rapid development of web-based applications that are platform independent. JSP technology separates the user interface from content generation enabling designers to change the overall page layout without altering the underlying dynamic content. JSP technology uses XML-like tags and scriptlets written in the Java programming language to encapsulate the logic that generates the content for the page. Additionally, the application logic can reside in server-based resources (such as Java Beans component architecture) that the page accesses with these tags and scriptlets. Any and all formatting (HTML or XML) tags are passed directly back to the response page. By separating the page logic from its design and display and supporting a reusable component-based design, JSP technology makes it faster and easier than ever to build web-based applications. Servlet engine Client’s browser EJB engine database Xsp Cocoon Admini strative tools EJBs Servlets & JSP Tomcat JBoss Oracle Application server Server side network Figure 30: INTELLECT's system architecture JSP technology is an extension of the Java Servlet technology. Servlets are platform-independent, 100% pure Java server-side modules that fit seamlessly into a web server framework and can be used to extend the capabilities of a web server with minimal overhead, maintenance, and support. Unlike other scripting languages, servlets involve no platform-specific consideration or modifications; they are Java application components that are downloaded, on demand, to the part of the system that needs them. Together, JSP technology and servlets provide an attractive alternative to other types of dynamic web scripting/programming that offers platform independence, enhanced performance, separation of logic from display, ease of administration, extensibility into the enterprise and most importantly, ease of use. Java Beans component architecture is the platform-neutral architecture for the Java application environment. It can develop or assemble network-aware solutions for heterogeneous hardware and operating system environments, within the enterprise or across the Internet. Java Beans component architecture extends "Write Once, Run Anywhere" capability to reusable component development. In fact, the Java Beans architecture takes interoperability a major step forward - every code runs on every operating system and also within any application environment. A beans developer secures a future in the emerging network software market without losing customers that use proprietary platforms, because Java Beans components interoperate with ActiveX. Java Beans architecture INT-D13FINAL © The INTELLECT consortium – 2002 Page 39 of 50 INTELLECT IST-1999-10375 connects via bridges into other component models such as ActiveX. Software components that use Java Beans APIs are thus portable to containers including Internet Explorer, Visual Basic, Microsoft Word, Lotus Notes, and others. The JavaBeans specification, which was completed ahead of schedule, defines a set of standard component software APIs for the Java platform. The specification was developed by Sun with a number of leading industry partners and was then refined based on broad general input from developers, customers, and end-users during a public review period. The Java Servlet technology provides web developers with a simple, consistent mechanism for extending the functionality of a web server and for accessing existing business systems. A servlet can almost be thought of as an applet that runs on the server side - without a face. Java servlets have made many web applications possible. Servlets are the Java platform technology of choice for extending and enhancing web servers. They provide a component-based, platform-independent method for building web-based applications, without the performance limitations of CGI programs. And unlike proprietary server extension mechanisms (such as the Netscape Server API or Apache modules), servlets are server- and platform-independent. This leaves you free to select a "best of breed" strategy for servers, platforms, and tools. Servlets have access to the entire family of Java APIs, including the JDBC API to access enterprise databases. Servlets can also access a library of HTTP-specific calls and receive all the benefits of the mature Java language, including portability, performance, reusability, and crash protection. Today, servlets are a popular choice for building interactive web applications. Third-party servlet containers are available for Apache Web Server, iPlanet Web Server, Microsoft IIS, and others. Servlet containers can also be integrated with web-enabled application servers, such as BEA WebLogic Application Server, IBM WebSphere, iPlanet Application Server, and others. JSP technology is an extension of the servlet technology created to support authoring of HTML and XML pages. It makes it easier to combine fixed or static template data with dynamic content. 7.2 Installation Guide of the client’s software The Java 3D API Version 1.2.1_02 Maintenance Release Implementation is available for download. The Java 3D 1.2 API is a set of classes for writing three-dimensional graphics applications and 3D applets. It gives developers high level constructs for creating and manipulating 3D geometry and for constructing the structures used in rendering that geometry. Remember, before installing Java 3D you will first need to install the Java 2 environment. In addition to several bugs fixed and better performance, the Java 3D 1.2.1 implementation also contains these additional features: • Documentation For Utility Classes • Example programs changed to use new Orbit Behavior classes Java 3D delivers Java's "write once, run anywhere" benefit to developers of 3D graphics applications. Java 3D is part of the JavaMedia suite of APIs, making it available on a wide range of platforms. It also integrates well with the Internet because applications and applets written using the Java 3D API have access to the entire set of Java classes. The Java 3D API draws its ideas from existing graphics APIs and from new technologies. Java 3D's low-level graphics constructs synthesize the best ideas found in low-level APIs such as Direct3D, OpenGL, QuickDraw3D, and XGL. Similarly, its higher-level constructs synthesize the best ideas found in several scene graph-based systems. Java 3D introduces some concepts not commonly considered part of the graphics environment, such as 3D spatial sound. Java 3D's sound capabilities help to provide a more immersive experience for the user. Java 3D was designed with several goals in mind. Chief among them is high performance. Several design decisions were made so that Java 3D implementations can deliver the highest level of performance to application users. In particular, when trade-offs were made, the alternative that benefited runtime execution was chosen. Other important Java 3D goals are to - Provide a rich set of features for creating interesting 3D worlds, tempered by the need to avoid nonessential or obscure features. Features that could be layered on top of Java 3D were not included. INT-D13FINAL © The INTELLECT consortium – 2002 Page 40 of 50 INTELLECT IST-1999-10375 - Provide a high-level object-oriented programming paradigm that enables developers to deploy sophisticated applications and applets rapidly. - Provide support for runtime loaders. This allows Java 3D to accommodate a wide variety of file formats, such as vendor-specific CAD formats, interchange formats, and VRML97. The installation guide for the client preparation will be published on the web-site of INTELLECT. If interested people want to play with the prototype, a direct link to our INTELLECT server is here available, which you can use. But keep in mind it is only a prototype, what means that we have to go on with the development of further features, embed more user requirements, and reach a more stable status as now. Therefore it is current only a beta version, which is maybe sometimes not available online. You need at first a 3D Java extension to see the 3D product and to be able to configure it. Therefore, the client must have installed: 1. Install the java plugin (with the java JRE 1.3.1): file j2re-1_3_1-win-i.exe (8,1 Mbyte) 2. Install the Java3D software (not a beta version): file java3d-1_2_1_01-win32-opengl-rt.exe (2,5 Mbyte) for Windows 98/NT/2000 (also a Solaris version is available) 3. Ensure that it does install in the right Virtual Machine (VM) the one that is used by the plug-in. (usually c:\program files\javasoft\jre) 4. Have only one Java Run Environment on your computer! 5. Add the vrml97.jar (317 Kbyte) to the classpath of the client. This can be done by several way either to the classpath in the autoexec.bat or in the /lib/ext directory of the right VM (ONLY AVAILABLE FOR BRUSSELS AND INTELLECT PROTOTYPE NOW!!!) After the short installation phase you can play with the different available prototypes! 7.3 Product Data Model Primary goal during the specification of the INTELLECT Product Data Model was flexibility in respect to the different kinds of products that can be categorised and simplicity of design in order to minimise the overhead imposed by the model itself. Product data can be entered into the INTELLECT database either manually through an application (see Figure 33) designed to be used by the knowledge engineer administering the knowledge base or in a bulk fashion i.e. from data exported from a backoffice. The structure consists of the following entities: 1. Pre-configured products 2. Products 3. Component types 4. Components 5. Component categories Products represent a package sold to the customer. This can be either a single object or a combination of components comprising a specific variant. Products contain one or more list of components. Lists are meant to support structuring of related components. An example of a product with more than one component list is kitchen furniture. The equipment for a whole kitchen could be specified as one product that is clearly structure in one or more related groups of components (i.e. a table and chairs, various cupboards, etc.). A computer on the other hand will most likely contain only one component list for all components. The configuration of the component lists and the categorisation of the components is up to the administrator’s discretion. INT-D13FINAL © The INTELLECT consortium – 2002 Page 41 of 50 INTELLECT IST-1999-10375 Hard Categories Soft Categories Component Types Components HardDisks Mainboards AVEquipment High-End Low-End IBM DCAS3440 Maxtor 4567 IBM DORS3240 U2W, 8 GB, 5400rpm U, 4 GB, 5400rpm UW, 2 GB, 5400rpm Figure 31: Component types, Components and Categories Products exist furthermore in one of two variations, pre-configured products and new products. Preconfigured products are the ones used to generate the product catalogue and are used as templates for generating new variants. Every time a user decides to configure a product, a copy of the preconfigure version there of is made. The new product is then modified to produce the desired variant. The differentiation between the two kinds of products is required in order to prevent modification of the original designs. Component types represent a certain type of part, possibly a specific model (i.e. the IBM DCAS2340 hard-disk). A component type is used to define the attributes related to this object. A dynamic attribute mechanism allows for arbitrary types and quantities of attributes. It is important to note that attributes of component types have not values, they simply provide the infrastructure needed for the usage of attributes in a specific product. Components on the other hand represent concrete elementary parts sold either separately or as part of a product. Such components are instances of a specific component type and contain concrete values for the defined attributes. A component for a specific type of hard disk (specific component type) i.e. would have concrete values for attributes like size, speed, form-factor etc. INT-D13FINAL © The INTELLECT consortium – 2002 Page 42 of 50 INTELLECT IST-1999-10375 <<Interface>> IProduct updateCurrentProductKey() playMultimedia() setPrice() getPriceOfAllComponents() getManufacturer() showDetails() PreconfiguredProduct createNewProduct() NewProduct customiseProduct() show3D() getComponentList() getPossibleComponents() changeComponent() ComponentCategoryConstraint Product ComponentCategory price : Integer availibility : Boolean productKey : Integer productName : String taxrate : Double 1 CategorySpecificData CategoryType : String playMultimedia() getPriceOfAllComponents() showDetails() setPrice() updateCurrentProductKey() getManufacturen() ComponentConstraint 1..* ComponentList Item ComponentType id : Integer multimediaUrl : String configurationContrains manufacturerName : String manufacturerId : Integer 1..* updateItem() deleteItem() getManufacturerName() getManufacturerLogo() showInfo() createItem() componentTypeName : String emptyComponentAttributes : HashTable IComponent getComponentTypeName() setComponentTypeName() componentKey : Integer componentName : String filledcomponentAttributes : HashTable componentPrice : Integer availibility : Boolean setAttributesType() getAttributesType() ComponentTypeattribute name : String valueType : String setName() setValueType() getName() getValueType() FilledComponentAttribute ComponentAttribute_Integer ComponentAttribute_Float ComponentAttribute_String attributeValue : Integer attributeValue : Float attributeValue : String setValue() setValue() setValue() Figure 32: INTELLECT product data model class-diagram INT-D13FINAL © The INTELLECT consortium – 2002 Page 43 of 50 INTELLECT IST-1999-10375 Figure 33: Tree view of a complete product Component categories exist in two variants hardware categories and soft categories. Hard categories represent a whole group of components that have the same basic function (i.e. all hard-disks), soft categories define configuration preferences and general criteria. Both types are used to reduce the number of components that are suitable. Soft categories are also used to define the “soft” criteria suitable for this product domain (see Figure 31). The knowledge engineer designing the knowledge base for a new Configurator has to decide what groups of components are necessary in order to provide the end-user with a sufficient set of options during configuration. PCI Card, Elsa Erazor Equality Constraint (=) ConnectionInput=PCI Mainboard, Asus xyz ConnectionOutput=PCI Figure 34: Constraints Constraints between attributes of specific component types enforce “construction” rules (see Figure 34). Constraints are operators that can be defined nearly arbitrarily and can be used to support new configuration behaviour. The Attributes “ConnectionImput” and “ConnectionOutput” together with the equality constraint i.e. are used to define which components fit together physically and what sort of interfaces each component type requires or makes available. Due to the modular and extensible design of the INTELLECT Configurator, it is relatively easy to add new Constraint types (i.e. bigger than, smaller than) that make new kinds of configurations possible. Products used for testing in the first version of the Configurator (scooters and computers) were fully modelled using only the equality constraint and appropriate Component Categories. INT-D13FINAL © The INTELLECT consortium – 2002 Page 44 of 50 INTELLECT IST-1999-10375 7.4 User Handling The VR module is the main Configurator user-interface. It aims to present the product in a way that is as realistic as possible. This is achieved through a product visualization in 3D and interactive navigation in 3D space. The user can interact with the product: turn, zoom, select and configure a component; the Configurator supplies the possible configurations. The 3D view of the products is obtained by assembling 3D models of its components stored in VRML files. This functionality is implemented with the help of two programmes: The manager application is intended for the shop operator, which will be in charge of the management of the 3D data, and the assembling rules definition. The user application is intended for customer use, and will mainly offer a 3D view of the product, which the customer will be able to view in several positions, and to configure according to his wishes. The use of Java3D (the 3D API for Java) allows the rendering of a 3D world in a Java application or a Java applet. Java3D provides a collection of high-level constructs for creating, manipulating and rendering 3D geometric objects in a virtual universe. The Java3D world corresponds to a scene graph, which is an arrangement of 3D objects in a tree structure that completely specifies the contents of the virtual universe, and the way it is rendered. VRML files can be loaded and added to a Java3D universe thanks to the VRML97 library, defined by the VRML-Java3D working group. 7.4.1 Managing the VR representation of products A product is composed of several components; each has a 3D model described in a VRML file. To view the product, we have to assemble correctly the components. To define the assembling rule between two components of a product, an “Anchor” is needed for each component. This anchor is composed of an origin and an orientation relative to the origin of the scene. When the two components are assembled, their two anchors correspond exactly. The two anchors for an assemblage have the same name. This mechanism corresponds to the “connectionInput” and “connectionOutput” attributes used by the Configurator. <<include>> AddNewAnchor ViewAnchor <<include>> EditAnchor <<communicate>> <<include>> Manager SelectAnchor <<communicate>> MoveAnchor Manager <<communicate>> SuppressAnchor Assemble SaveComponent Figure 35: Use Case INT-D13FINAL © The INTELLECT consortium – 2002 Page 45 of 50 INTELLECT IST-1999-10375 The shop operator has to define the two anchors for each assemblage in the manager application, and then the assemblage can be calculated by composing the transformations corresponding to the anchors. As shown on Figure 35, the manager (the shop operator) has the possibility to load a single component, and manage its anchors. When the manager loads a component, the application shows its 3D model and its existing anchors in the Java3D view (ViewAnchor). The anchors are also displayed in a list. The manager can select an existing anchor in the list, or in the visualisation area (SelectAnchor), but only if he is in the interaction mode (see Navigation); the selected anchor appears in a different colour than the others; it can be edited in a separate window (EditAnchor). While an anchor is edited, its values can be modified in the edition window or by “drag and drop” in the visualisation area (MoveAnchor), but in order to avoid confusions, the manager is not allowed to select another anchor until the edition window has been closed. The manager also has the possibility to create a new anchor (AddNewAnchor), which is instantiated with default values. He can also suppress an existing anchor (SuppressAnchor). Figure 36: Anchors for the assemblage between a frame and a wheel To choose the name of an anchor, the manager will have to select the second component of the assemblage from the existing components and then choose one of its anchors, if it has already been created, or define a new name if the anchor does not exists yet on the other component. The manager will have to save the component (SaveComponent) if he wants its modifications to be taken in account. He if does not, and tries to exit the application, a dialog frame will ask him for confirmation, and give him the choice to save or not. The manager has the possibility to view the results of the assemblage (whether the component has been saved or not) thanks to the Assemble functionality. He will have to choose from a list of available components, including the one which is edited, the components he wants to assemble. The application will load every component, read the anchors list and calculate the corresponding assemblage. Once all the anchors needed for a product have been placed, the manager will ask the configuration module if its product is correct (to be sure he hasn’t forget a component or an anchor), if yes, he has the possibility to save this product as a pre-defined product. The navigation possibilities are the same in the customer and manager applications. Navigation includes everything that allows the user to see the object (product or component) from a different point of view. Some usual navigation possibilities in Virtual Reality applications are more or less userfriendly, depending on the viewed scene. Navigation using the keyboard for example (←, ↑, →, ↓) are mostly used in games or worlds the user can walk into; but in our application this navigation type become very unfriendly; turning around an object is very difficult, and it is better way for the user to manipulate (translate or rotate) directly the object, and make it move (MoveObject). INT-D13FINAL © The INTELLECT consortium – 2002 Page 46 of 50 INTELLECT IST-1999-10375 Figure 37: VR Manager Application 7.4.2 VR Graphical User Interface for the customer The customer sees in the main window the pre-defined product he has chosen in the eShop, and is able to interact with it and manipulate it in various ways. Three modes are available for the user: translate, rotate and interact. The translate mode allows the user to use the mouse to translate the object in the current visualisation plan, the rotation mode allows the user to rotate the object with the mouse, and the interaction mode disables navigation with the mouse (but not the other navigation means), so that the user can use the mouse to interact with the scene (select an anchor or a component, or move an anchor). To make it easier for the user who may not be used to use the mouse for moving objects, the application calculates a list of predefined ones (Face, Back, Right, Left, Up, Down), using rotations from the initial viewpoint. Another viewpoint allows the user to see all the objects of the scene. The user also has the possibility to zoom in or out with the zoom functionality. He can give directly the zoom value, increment or decrement the existing one, or reset the zoom. All these navigation possibilities allow the user to view the object in several points of view even if he isn’t very used to navigation, and to move objects with the mouse, which may seem difficult for a novice, but which bring the sensation of manipulating the object. INT-D13FINAL © The INTELLECT consortium – 2002 Page 47 of 50 INTELLECT IST-1999-10375 8 Conclusion After the first pilot has been started the modules were improved continuously. That means new user requirements did arise during the pilot phase and must fulfil from the developers, handling limitations need a new adaptation, missing features had to extend the prototype, and a stable platform had to created for testing and evaluation. Nevertheless there are still some points missing regarding the modules after the pilot phases, like 1. eShop: SSL integration, creating of XML pages for HiTEC, new backend functionality regarding the end-users requirements, and an administration tool 2. Virtual reality: Integration of the 3D window into the web surface, an avatar for animations, 2D visualisation into the configurator, and producing 3D data 3. Configuration: 3D data is missing for the configuration database and some input for the database at all 4. Help-desk: There were no mirroring available for the pilots, because of Tomcat 4.x had to be integrated into JBoss, but Tomcat 4.x was only available since the end of the year of 2001. 5. 3D data: We need the 3D data for the configuration. That means problems e.g. for Interset because they really don't need it and in addition the implementation effort at the client computer is too high at the moment. The other end-users has the same problems. Only HiTEC needs 3D visualisation really. And they didn’t get this data it from their producer. 6. Product presentation: The 2D configuration was selected for the pilots of Interset and Blauwerk. But the integration into the 3D module and into the configurator is still missing. In the meantime, the prototypes are stable enough to seriously invite customers to present the system to them now. But the project partners still need additional work of the different modules after the project’s end if they want to create a real product. INTELLECT had set-up a list of maintenance points concerning all modules. That means requirements for up-dates, bug fixes, additional requirements and so on. The development partners of INTELLECT used the pilot phase to collect new experience about their developed system. Also they were in contact with the end-users continuously to find out limits, usability, efficient handling, and problems. After all they got a stable prototype which consist of different modules which are very flexible. The end-users got a big support from developers and set-up their new designed web-pages with new content and features. They will support the product also after the projects lifecycle to bring it into the market and present real worked demo platforms. Against all missing features, the project INTELLECT was a success story for all partners within the project. INT-D13FINAL © The INTELLECT consortium – 2002 Page 48 of 50 INTELLECT IST-1999-10375 9 Annex 9.1 Pictures from the whiteboard from Blauwerk’s prototype Figure 38: eShop structure (navigation) with the main interest of products INT-D13FINAL © The INTELLECT consortium – 2002 Page 49 of 50 INTELLECT IST-1999-10375 Figure 39: Product site of a choosen scooter with applet window and video streaming Figure 40: 3D visualisation with visible configuration, actual prices, and animation Figure 41: The other web sites without products INT-D13FINAL © The INTELLECT consortium – 2002 Page 50 of 50
Similar documents
Deliverable (public)
1.2 INTELLECT Project Objectives To improve this situation the goal of the proposed project is the creation of an online configuration tool
More information