Towards a Theory of Conceptual Modelling

Transcription

Towards a Theory of Conceptual Modelling
Towards a Theory of Conceptual Modelling
ETheCoM 2009, invited talk
Nov. 09, 2009
Bernhard Thalheim
Dr. rer.nat.habil.
Prof. @ Christian Albrechts University at Kiel, Germany
Department of Computer Science
Information Systems Engineering Group
(∗)
Kolmogorov Professor h.c. @ Lomonossov University Moscov, Russia
Today:
Information systems development is
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
• development in the large
taught almost nowhere
• however everybody likes his/her languages and his/her understanding - typically incompatible with everybody else
• might be based on database components and pattern of
schemata
¨
¥
Our Background
§
¦
• schema library consisting of more than 10.000 applications
largest with 95.000 types
• used for learning experiences from developers
• evolution, change, re-engineering of most applications at least after
6 month of existence
• migration, integration and collaboration projects
Why this talk?
¨
¥
Is there any theory or technology of conceptual modelling?
§
¦
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
Garcia-Molina, Ullman, Widom: ER/relational/other data models, information integration
Hoberman: elements (entity, data elements, relationships), definition, subject area, logical model, physical model, best approach,
validate
Mannila, Räihä: ER/relational/other data models, design principles, constraints, properties of relational schemas, mappings, transformations
Olivé: elements (entity types, relationship types, constraints, special
relationship types), reification, generic relationship types, derived
types, taxonomies, domain events, action request events, UML diagrams, metamodeling, MOF
Simsion: overview (what is modeling, normalisation, ER, subtypes/supertypes), advanced data modeling (constraints, normalisation, identity, attributes, time dimension, business rules, corporate
data modelling
Large and Complex Schemata: MDA
CIDOC
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
Large and Complex Schemata: MDA
CIDOC Class Hierarchy
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
Large and Complex Schemata: Starting
with “Simple” Schema and Evolving
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
Large and Complex Schemata: Evolved and
Problematic Schema
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
Seven Myths of Formal Methods
¨
¥
Anthony Hall
§
¦
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
(1) Formal methods can guarantee that software is perfect.
(2) Formal methods are all about program proving.
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
(3) Formal methods are only useful for safety-critical systems.
(4) Formal methods require highly trained mathematicians.
(5) Formal methods increase the cost of development.
(6) Formal methods are unacceptable to users.
Information
Concept
Topic
c
°B.
Thalheim
(7) Formal methods are not used on real, large scale software.
Seven + Seven Myths of Formal Methods
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
¨
¥
Dines Bjørner
§
¦
(8) Formal Methods delay the development process.
(9) Formal Methods are not supported by tools.
(10) Formal Methods mean forsaking traditional engineering design
methods.
(11) Formal Methods only apply to software.
(12) Formal Methods are not required.
Content
Information
Concept
(13) Formal Methods are not supported.
Topic
c
°B.
Thalheim
(14) Formal Methods people always use Formal Methods.
Myths of Modelling
(1) Modelling equals documentation
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
(2) You can think everything through from the start
(3) Modelling implies a heavyweight software process
(4) You must “freeze” requirements
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
(5) Your design is carved in stone
(6) You must use a CASE tool
(7) Modelling is a waste of time
(8) The world revolves around data modelling
(9) All developers know how to model
(10) Modelling is independent on the language
¥
¨
Let’s discuss the last point in detail
§
¦
Wrong or Inappropriate Language: Traffic
Control based on Petri Nets
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Simple but classical solution
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
Picture becomes more complex if time control is incorporated
N/S red
W/E green
:
Time for switch
Y
¾
N/S green
¼
: y
s
z
j
9
Time for wait
ª
j
W/E red
Wrong or Inappropriate Language: Look at
Reality
• Situation in reality
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
• Neighbourhood street crossing
must be considered
• Complex time management
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
• Different kinds of traffic: pedestrians, tram, bus
Wrong or Inappropriate Language: Traffic
Control based on Petri Nets
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
Wrong or Inappropriate Language: Traffic
Control based on State Charts instead on
Petri Nets
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
Local-as-View Modelling: Traffic Control
based on ASM and Conceptual Modelling
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
Global states: NSredEWgreen, NSredEWred, NSgreenEWred
optionally: NSredEWred, NSnothingEWnothing
avoiding conflict states such as NSgreenEWgreen
Abstract state machine rules:
Controller
location
state
clock
reset
switch
e
...
...
...
...
...
if Switch(e) then Update(e,collocated(e)); ChangeSwitch(e)
ChangeAction := getState; choosePossibleStateChange(state);
apply(possibleStateChange(state))
AlarmAction := on alarm changeStateToErrorState
Clock := on tick observeWhetherChangeRequired
NormalAction := if change = true then ChangeAction
PedestrianCall := on callAtPoint(cp)
ChangeNextStepIssuedAt(cp)
Local-as-view Modelling: Traffic Control
based on ASM and Conceptual Modelling
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Global states: NSredEWgreen, NSredEWred, NSgreenEWred
Abstract state machine rules
Database as the backend machine
State
Kind
¾
KindOf
State
-
Global
LightState
¾
¾
State
Change
Concept
Topic
c
°B.
Thalheim
State
Change
Kind
6
Timer
Schedule
Strategy
¾
NextStepIssuedAt
Request
Change
Request
L
- Pedestrian
CallPoint
Content
Information
-
∗
local-as-view: πEW
(GlobalLightState) for EW direction
Formalisation:
Ten Commandments
¨
¥
Jonathan P. Bowen & Michael G. Hinchey
§
¦
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
(10)
Thou
Thou
Thou
Thou
Thou
Thou
Thou
Thou
Thou
Thou
shalt
shalt
shalt
shalt
shalt
shalt
shalt
shalt
shalt
shalt
choose an appropriate notation.
formalise but not overformalise.
estimate costs.
have a formal methods guru on call.
not abandon thy traditional development methods.
document sufficiently.
not compromise thy quality standards.
not be dogmatic.
test, test, and test again.
reuse.
Another three?
(1) Thou shall meet intentions of developers, ...
(2) Thou shall provide a usable notation, i.e. for verification, validation,
explanation, elaboration, and evolution.
(3) Thou shall be robust against misinterpretation, errors, ...
Formalisation: Kinds and Needs
Formal development: a software development which uses one or more formal techniques
and it may then use these in a spectrum.
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
Formal software Development: a software development in which specifications are
expressed in a language with a formal syntax so that all specifications can be judged wellformed or not, with a formal semantics so that all well-formed specifications have a precise
meaning, and a (relatively complete) proof system such that one may be able to reason over
properties of specifications or steps of formally specified developments from a more abstract
to a more concrete step.
Additionally a formal technique may be a calculus which allows developers to calculate, to
refine “next”, formally specified development steps from a preceding, formally specified step.
Systematic (formal) development: a software development which formally specifies
whenever something is specified, but which does not (at least only at most in a minor of
cases) reason formally over steps of development.
Rigorous (formal) development: a software development which formally specifies
whenever something is specified, and which formally express (some, if not all) properties
that ought be expressed, but which does not (at least only at most in a minor number of
cases) reason formally over steps of development, that is, verify these to hold, either by
theorem proving, or by model checking, or by formally based tests.
Formal (formal) development: a software development which formally specifies whenever something is specified, which formally expresses (most, if not all) properties that ought
be expressed, and which formally verifies these to hold, either by theorem proving, or by model
checking, or by formally based tests.
Main Dimensions of Conceptual Modelling
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
(1) Modelling language constructs are applied during conceptual modelling. Their syntactics, semantics and pragmatics must be well
understood.
¨
¥
30% coverage in most modelling methodologies because of “iron” selection
§
¦
(2) Application domain gathering allows to understand the problems to
be solved, the opportunities of solutions for a system, and the requirements and architecture that might be prescribed for the solution
that has been chosen.
¨
¥
60 % coverage in most modelling methodologies because of pressure, ...
§
¦
(3) Engineering is oriented towards encapsulation of experiences with
design problems pared down to a manageable scale.
¨
¥
10 % coverage in most modelling methodologies because ????
§
¦
Application Domain Dimension
¨
¥
Everybody develops his/her schema, his/her implementation
§
¦
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
Application domain development based on “fly by”
• mostly by domain workers without IS insight
• scope of current (and not future) interest
• point of view of subject (subject-oriented)
• tolerance level of users
Flat schemata based on
• binarisation,
• without views, and
• without internal computational support
Near-sighted reflection of ongoing demand
• data needed for analysis but not at the right granularity, precision
• data cleansing is of less importance
• “my-data-is-my-personal choice”
The Neglected Dimension: Engineering
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
¨
ingenerare, “to create”
§
¥
and/or
“to contrive”
¦
Sufficient quality instead of optimal quality defined on the
purpose
e.g. integrity constraints that are really necessary
Living with errors as long as users can live with them, living with
enforcement and fetching modifications in time
Living with deficiencies of technology, e.g. missing support for full
storage (sliding window techniques)
Providing forgetful data handling with automatic background
archiving or deletion or distribution
User-demand driven query answering, search drill-down, eager/lazy enforcement, data granularity, variety of views depending
on task/profile/role under consideration
Engineering
¨
ingenerare, “to create”
§
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
and/or
¥
“to contrive”
¦
application of science to the optimum conversion
of the resources of nature to the uses of humankind
creative application of “scientific principles to design or develop structures, machines, apparatus,
or manufacturing processes, or works utilizing them singly or in combination; or to
construct or operate the same with full cognizance of their design; or to forecast
their behaviour under specific operating conditions; all as respects an intended function,
economics of operation and safety to life and property”
manufacture or assembly of engines, machine tools, and machine parts
engineering: military - civil - mechanical - chemical - industrial
great body of special knowledge
involves extensive training in the application of that knowledge
standards of engineering practice
functions: (scientist; to know verified, systematized knowledge of the physical world),
(engineer; to do and bring knowledge to bear on practical problems)
engineer is not free to select the problem that interests him, solves problems as they arise
solution must satisfy conflicting requirements
(technical, technological, economical, ..., social)
types of resources: materials, information and energy
Treatment
of errors
¨
¥
be aware and develop mechanisms
§
¦
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Systemic errors e.g. based on data exploration, data re-usage, data merging,
falsifications, biases, wrong models
Systematic errors e.g. due to abstractions, restrictions in accessible data, dirty
data, approximations, computations
Stochastic errors e.g. based on assumptions for occurrence of errors, their distribution functions and their contribution within the model and to the variables
Handling by
error prevention with direct correction, duplicate extraction and elaboration,
validation, and verification
(probabilistic) error models with maximal error, error spreading and multiplication, confidence intervals, means, outlier detection, time series abstractions
error cleansing algorithms with(out) data recovery
Content
Information
Concept
Topic
c
°B.
Thalheim
Development Dimensions of Conceptual
Modelling
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
(0,n)
(0,n)
(0,1)
Aspect
Partner
Introduction
performs
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
restricted to
(0,n)
(0,1)
Activity
based on
(0,1)
Resource
(0,1)
uses
creates/revises
(0,1)
Work product
Content
(0,n)
Information
Concept
Topic
c
°B.
Thalheim
(0,n)
¨
¥
complexity # common misbelief in existence of a theory of CM
§
¦
The Work Product Development
Dimension
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
¨
¥
The main product is the model; it shall however used for realisation!
§
¦
The work product
depends on abstraction layer e.g. requirements, specification, realisation or implementation layer,
depends on granularity and precision of the work product itself,
depends on resources used for development such as the languages,
depends on level of separation of concern such
as
static/dynamic properties, local/global scope, facets,
depends on quality properties of the input , e.g. requirements,
completeness, conciseness, coherence, understandability,
depends on decomposition of the work products in ensembles of
sub-products, and
satisfies quality characteristics
quality in use
internal quality
external quality
The Creator/Actor Development
Dimension
¨
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
¥
Everybody is different and thus modelling has different results
§
¦
Attitude and preferences
Ability to understand, to model, to reason, to survey, to communicate with others, to analyse, to construct systems, to validate or
to verify or to test models, to use or develop documentations
Master complexity, improvements, realisation
Knowledge, skills, competency for representing world, for coping
with representations
Restricted expressivity due to restricted leads to preferring local
reasoning
Experience to cope with varieties of problem solutions through
generic problem solving
Referential solutions to be used for solution of similar problems
together with refinement of the given approach
The Activity Development Dimension
¨
¥
The task is never complete: p´tamoi rh´ousi ‘the rivers flow’
§
¦
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
Scope insight for conscious handling of restriction, capabilities, opportunities;
Guiding rules for convenience, for completion, refinement, and extension;
Development plans for partial delivery of work products, partial
usage and deployment;
Theories supporting development of models;
Quality characteristics for model completion, model evolution,
model engineering
Mappings styles for mapping among abstraction layers
The Partner Development Dimension
¨
¥
Models tend to be too large for a singleton person!
§
¦
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
Roles during activities such as stakeholder, developer, consultant,
supplier, contractor, documentation developers, or finally business
user
Collaboration partnership based on communication acts, cooperation business processes, and coordination agreements
Teamwork during all activities with separation of different tasks;
Historical people such as teachers, legacy (better heritage) developers, coders, ...
Builders of their products based on the current solution
Summarising: The Gaps of Conceptual
Modelling
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
“Partial reality”
Part
of reality
6
Things of
reality
Concept
Topic
c
°B.
Thalheim
Observed
property
- Predicator
“Topic”
6
Modelling
decision
Foundation of
decisions
6
6
Usage of
theory
Content
Information
¾
?
Modality
Exactness
®
6
Revision
during the
development process
“Schema” as result
and partial point of view
of a database
development process
Confidence
Context
6
U
acts
within
?
Modeler
¾
under
usage
?
Reference
model
Modelling
Arts: working of plastic materials by hand to build up form
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
Mathematics: method of scientific investigation of systems
no necessity to construct an actual physical model of the system
mathematical model: a description of the system in some algorithmic language
divided into individual parts and the state of each part described by some
system of parameters
description of the relationships between the separate parts
Cybernetic systems: self-improvement, self-teaching and self-development
to model
• to plan or form after a pattern or shape
• to make into an organization (as an army, government, or parish)
• to produce a representation or simulation to model a problem
• to construct or fashion in imitation of a particular model
General Principles and Properties of
Modelling
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
¨
¥
(D, C, (R, ρ, θ, Ψ), G, W)
§
¦
• things D under consideration
• concepts C described in a language L
• relationship R ⊆ D × C with
• restrictions ρ to its applicability
• modality θ or rigidity of the relationship
• confidence Ψ in the relationship
• agreed upon within a group G within a culture C
• valid in a certain world W
¨
¥
Principle of Abstraction
§
¦
we may however ‘invent’ concepts (distortion) and ‘restrict’ the model
(extension of properties)
General Principles and Properties of
Modelling
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
¨
¥
(D, C, (R, ρ, θ, Ψ), G, W)
§
¦
mapping property: have an original,
truncation property: the model lacks some of the ascriptions made to the original,
pragmatic property: the model use is only justified for particular
model users, tools of investigation, and period of time,
extension property: represent judgments which are not observed
for the originals
distortion property: for improving the physical world or for inclusion of visions of better reality
Content
Information
Concept
Topic
c
°B.
Thalheim
Stachowiak: 1-3
Qualities
of a Model¥
¨
Mainly Qualities of Use
§
¦
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
Static qualities for a model
Development quality: pervasiveness, analysability, changeability, stability, testability, privacy of the models, ubiquity
Internal quality: accuracy, suitability, interoperability, robustness, self-contained, independence
Quality of use: understandability, learnability, operability, attractiveness, appropriatedness
Dynamic qualities within a selected development approach
executability, refinement quality, scope restriction, effect preservation, context explicity, completion tracking
modelling properties: monotonicity, incrementality, ...
see below
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
Modelling Ingredients = Languages +
Restrictions + Negotiation +
Methodologies
Languages with syntactics, semantics, pragmatics
inductive expression formation based on alphabets
behaviour defined on expressions
Restrictions depending on logics
deontic, epistemic, modal, belief, preferences
shortcuts, ambiguities, ellipses
inherent language semantics
classes of constraints versus sets of real-life constraint
Negotiation by identification and analysis of the barriers
(strategic, psychological, legal, and structural)
to management or resolution of conflict
development of strategies to overcome these barriers
Methodology of development based on pragmatism, paradigms
Variables in modelling
¨
¥
e.g., static, deterministic, discrete, complex, intertwined
§
¦
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Evolution of the system
• Kind of system and its equations
Behaviour of variables
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
static dynamic
• Predictability of behaviour
deterministic stochastic
Values for variables
• Value domain
discrete continuous
Dimension of variables
• Complexity of model
one-dimensional complex, many-dimensional
Content
Information
Concept
Topic
c
°B.
Thalheim
Dependence among variables
• Correlation within model
independent correlated
Collaboration Based on a Commonly
Agreed
and Understood Language ¥
¨
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
Forget about religious wars about the language choice
§
¦
Inherent complexity of languages with pattern-based transformation to realisation languages
Ambiguity of expressions requires context extraction and injection;
variety of meanings as modularisation, context abstraction and privacy protection
Choice among (partially) equivalent expressions depending
on emphasis, appropriate appearance reflecting reality within the
understanding of the user community
¨
¥
Variety of models depending on preferences, users, usages, ...
§
¦
Partially defined semantics for economy of utterances
Language-dependent culture of representation
Capacity and appropriatedness of languages for certain aspects
within the application
¨
¥
Model ensembles, model suites, GLaV model sets
§
¦
The Sapir-Whorf Syndrome
B.L. Whorf, Lost generation theories of mind, language, and religion.
Ann Arbor, Mich., Popular Culture Association,
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
University Microfilms International, 1980.
D. Sapire, General causation, Synthése, 1991, 86, 3, 321–347
“Principle of linguistic relativity”: actors skilled in a language
may not have a (deep) understanding of some concepts of other
languages
The design and development quality depends on main success factors:
• structuring of the process itself,
• culture of people involved,
• skills of actors, and
• process capabilities
Language
Layers of Specification
¨
¥
Specification Engineering: Separation and Abstraction
§
¦
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
(1) Declaration layer based on logical formulas or constructs description
(2) Technical layer, e.g., methods for maintenance, by rules for compensation,
enactment strategies, auxiliary methods
operational semantics
(3) Technological layer under explicit consideration of implementation and refinement context
application
(4) Organizational layer by integration into the architecture of the system, by
obligations for users and for components of the system
establishment
(5) Economical layer: (economical and technological) feasibility, quality satisfaction
(6) Handling satisfaction of properties and predicting changes of satisfaction
(7) Optimisation for evolution and adaptation
(8) Experiences utilisation for innovation and generalisation
The Modelling Act
¨
¥
similar to the speech act
§
¦
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
(1) a selection and construction of an appropriate model depending
on the task and purpose and depending on the properties we are
targeting and the context of the intended system and thus of the
language appropriate for the system,
(2) a workmanship on the model for detection of additional information
about the original and of improved model,
(3) an analogy conclusion or other derivations on the model and its
relationship to the real world, and
(4) a preparation of the model for its use in systems, to future evolution
and to change.
Constituents of Modelling
¨
¥
Development of decisions and strategies
§
¦
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Aims of the integrated model and target outcome
criteria of fitting
similarity models
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
Integration and mapping of models depending
on their correlation and cohesion
Detection, description, experimentation of laws with
players, stability factors, entailers
main
Evolution of model ecosystems with explicit extraction of control parameters that effect or drive evolution
Establishing a Model
¨
¥
§Phases ¦
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
• Clarification phase
• Model construction phase
• Model experimentation phase, e.g. based on simulation methods
• Model optimisation phase
• Model validation phase
• Model application phase e.g. for decision making
Supporting Facilities
Blackboard for explicit treatment of open issues to be resolved
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
Obligation handler for explicit intermediate storage of questions/activities left open at a certain activity for resolution at a later
activity
Decision tracker for explicit recording of decision made
Alternative schemata for later revision of a schema
Schema equivalence recorder for later choice of another schema
that is is equivalent to the current one
Incompleteness manager for explicit incomplete elements in a model
Primitive Activities
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
according to kinds of data abstraction
(classification, instantiation; aggregation, decomposition;
generalization, specialization)
Operations: selection / union,
partition / union,
nest / unnest,
projection / product (or join)
Composition constructor C, new name t, types t1 , ..., tn
compose(t1 , ..., tn , C, t)
Decomposition for a type t, terms e0 , e1 , ..., en on t, names N0 , ..., Nn
decompose(t, e0 , e1 , ..., en , N0 , N1 , ..., Nn )
Extension
t00 = extent(t, t0 ) extend t by t0 in component c
Information
Concept
Topic
c
°B.
Thalheim
Initialization
generate(t) for abstract t
special composition
Computational Refinement
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
Given two abstractions M, M ∗ , refinement is based on
refinement of states
states of interest S, S ∗ , correspondence between the states of interest
abstract computation segments τ1 , ...., τm on M and σ1 , ...., σn on M ∗
(m,n)-refinement
locations of interest
equivalence relation ≡ on locations of interest
M ∗ is a correct refinement of M if
there for each M ∗ -run S0∗ , ...., Sk∗ , ... there is an M-run and sequences i0 < i1 < ....
and j0 < j1 < ... such that i0 = j0 = 0 Sik ≡ Sj∗k for each k and either
• both runs terminate and their final states are the last pair of equivalent states,
or
• both runs and both sequences are infinite.
Complete refinement: M correct refinement of M ∗ and M ∗ correct refinement
of M
Kinds of Refinement
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
Inside-out refinement: Inside-out refinement uses the current IS for extending
it by additional part. These parts are hocked onto the current specification
without changing it.
Top-down refinement: Top-down refinement uses decomposition of functions
in the vocabulary and refinement of rules. Additionally, the IS may be extended
by functions and rules that are not yet considered.
Bottom-up refinement: Bottom-up refinement uses composition and generalisation of functions and of rules to more general or complex. Bottom-up refinement also uses generation of new functions and rules that are not yet
considered.
Modular refinement: Modular refinement is based on parqueting of applications and separation of concern. Refinement is only applied to one module and
does not affect others. Modules may also be decomposed.
Mixed skeleton-driven refinement: Mixed refinement is a combination of refinement techniques. It uses a skeleton of the application or a draft of the
architecture. This draft is used for deriving plans for refinement. Each component or module is developed on its own based on top-down or bottom-up
refinement.
Generic Refinement Steps and Their
Correctness
Refinement pattern
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
Perspectives
and styles
-
?
Derivation of
generic
refinement
steps
¾
Development
contract
?
Generic refinement step
Consistency
conditions
-
?
Derivation of
specific
refinement
steps
?
Refinement step
¾
DBMS specification
assumptions
Often Neglected Activities
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
• Validation and verification activities
proof of properties, correctness, completeness
• Development of coherent constraint sets
• Detection of concepts and conceptions
• Quality management activities
e.g., robustness, modality, completeness, error forgiveness
• Transformation activities
compilation, interpretation, mapping
collaboration
abstraction
• Representing causal dependencies
The¨ Notion of the Model
¥
§Different flavours of the notion of a model ¦
(1) A system of postulates, data, and inferences presented as a mathematical description of an entity or state
of affairs.
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Encyclopedia Britannica
(2) Graphical, mathematical (symbolic), physical, or verbal representation or simplified version of a concept,
phenomenon, relationship, structure, system, or an aspect of the real world. The objectives of a model
include
(1) to facilitate understanding by eliminating unnecessary components,
(2) to aid in decision making by simulating ’what if’ scenarios,
(3) to explain, control, and predict events on the basis of past observations.
Since most objects and phenomenon are very complicated (have numerous parts ) and much too complex
(parts have dense interconnections) to be comprehended in their entirety, a model contains only those
features that are of primary importance to the model maker’s purpose. Models range from simple sketches
to computer programs with millions of lines of code, but all of them have one thing in common: some
elements of the actual ’thing’ are abstracted or mapped into the model. Models are divided into three
classes on the basis of their degree of abstraction.
• (1) Iconic model: least abstract , physical, ’look-alike’ model, such as a model airplane or train .
• (2) Analogous model: more abstract but having some resemblance to what it represents, such as a
chart, graph, map, network diagram .
• (3) Symbolic model: most-abstract model with no resemblance but only an approximation to what
it represents, such as a mathematical equation or formula
BusinessDictionary.com
(3) A schematic description of a system, theory, or phenomenon that accounts for its known or inferred
properties and may be used for further study of its characteristics: a model of generative grammar; a
model of an atom; an economic model.
Collins English Dictionary
Information
Concept
Topic
c
°B.
Thalheim
(4) A systematic description of an object or phenomenon that shares important characteristics with the object
or phenomenon. Scientific models can be material, visual, mathematical, or computational and are often
used in the construction of scientific theories.
Mariam-Webster Nr. 12
The Model
· is based on an analogy of structuring, functionality, or behaviour,
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
· considers certain application purposes, and
· provides a simple handling or service or consideration of the things
under consideration.
Model capacity:
· the model provides some understanding of the original;
· the model provides an explanation of demonstration through auxiliary
information and thus makes original subject easier or better to understand;
· the model provides an indication and facilities for making properties viewable;
· the model allows to provide variations and support optimisation;
· the model support verification of hypotheses within a limited scope;
· the model supports construction of technical artifacts;
· the model supports control of things in reality;
· the model allows a replacement of things of reality and acts
as a mediating means.
¨
¥
Choose whatever you like and master!?!?!?
§
¦
Explicit Treatment of Model Aims
¨
¥
Models have their aims and intends that should be separated
§
¦
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
Perception support for understanding the application domain
Explanation and demonstration for understanding
Preparation to management and handling of the original
Optimisation of the application domain operating
Hypothesis verification through the model
Construction of an artifact
Control of parts of the application
Substitution for a part of the application
¨
¥
Depending of the aim we shall use different schemata!
§
¦
Construction Principles
¨
¥
§Constructors for main structures and escort structures ¦
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
principles of constructing complex structures
meta
context
constructors
processing context
special
implementation
utilisation
associations
bulk
hierarchies
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
utilisation
quality
based on
based on
constructor
component hinge
type
star snowflake
history
layers
log ... strategic tactical
security performance
potential actual product ... multiset
classification
semantics
Explicit Consideration of Model Purpose
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
Construction purpose for construction of a solution to application
domain problems (either as business system or as embedded system)
Communication purpose among stakeholders
Analysis purpose for validation, verification, tests
Examination ad check purpose for application domain or constructed system
Documentation purpose for logging development decisions, alternatives, neglected parts, variants, reference models
Master complexity, improvement, evolution, and realisation
¥
¨
Each purpose requires its constructions and approaches!
§
¦
Componentisation
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
WorkTask
WorkTask
Assignment
q
1
Party
)
¾
i
PartyAsset
Assignment
PartyAsset
Assignment
Requirement
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
N?
°
Work
±6
M
Billing
Time Entry
)
¾
i
Fixed Asset
Assignment
Fixed Asset
Assignment
Inventory
Assignment
Inventory
Requirement
q
- Invoice
1
N?
°
q
- Asset
1
General Principles
¨
¥
Well-known principles but missing theory
§
¦
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
Modularisation: based on relatively independent component or subject architectures, orthogonalisation, parametrisation, based on Parnas information hiding and Liskov substitutability principles
with interfaces, obligations, views, hierarchies, responsibilities, integration, dependences, independence, incompleteness and fuzziness,
open world potential composability
data, functional and control decomposition
skeleton or crystallisation techniques
Abstraction: component or construction abstraction, context abstraction, implementation abstraction
pattern techniques and refinement, meta-structures, views
Coupling: easy/lazy enforcement, tight or loose coupling, integration
or collaboration
Principles
¤
¡
Esko Marjomaa: “Peircean” Reorganization in Conceptual Modeling Terminology
£
¢
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
Conceptualization principle: Only conceptual aspects of the application domain
should be taken into account when constructing the conceptual schema.
95% -principle: All the relevant aspects of the application domain should be described in the conceptual schema. instead of 100% principle; based on engineering insight
Formalization principle: Conceptual schemata should be formalisable in order to
be implementable.
Semiotic principle: Conceptual schemata should be easily interpretable and understandable.
Correspondence condition for knowledge representation: The modellens
should be such that the recognizable constituents of it have a one-to-one correspondence to the relevant constituents of the modellum.
Invariance principle: Conceptual schema should be constructed on the basis of
such entities found in the application domain that are invariant during certain
time periods within the application area.
Sub-schemata principle: In order to construct a good conceptual schema it is important first to construct relevant sub-schemata and then to search for connections
between them.
Choices for Specification: Modularisation
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
¨
¥
Implementation and localisation abstraction
§
¦
data, functional, and control decomposition
Explicit modularisation: skeleton-based modelling and architecturing
interfaces and collaboration of components
side-effect free computation and controlled effect on partners
Implicit modularisation: through name spaces
¨
¥
Advantages
§
¦
separation of concerns, discovery of basic concepts, validation and verification of development, efficiency of tool support, scoped changes,
evolution and extension, analysability, conservativeness, incrementality,
testability
¥
¨
Disadvantages
§
¦
does not support agile development and brute-force prototyping
Choices for Development Alternatives
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
Object-expanded or class-separated depending on the modelling language,
e.g. XML choices
• Venetian blind (full expansion of objects)
• Russian doll (DTD style)
• Salami slice (similar to ER diagrams)
Development strategy as 3-dimensional decision chart
Development direction (B,J,T)
6
UCT
ICT
UVT
IVT
UCJ
UMT
UMJ
UVB
UMB
ICJ
IMT
UVJ
IVJ
UCB
IMJ
B: bottom-up
T: top-down
J: jojo
Controlled
ICB
- development
(U,I)
UVB
M: modular
V: view-based
C: central
IMB
ª
Modularisation (C,V,M)
Dividing ridge for object (entity) types
Limiting expressiveness e.g., binarisation
U: uncontrolled
I: inside-out
Choices for Specification: Perspectives and
Styles
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
Structure-oriented perspective: structural description of the IS
+ semantic perspective
Behavior-oriented perspective: behavior of the IS during its lifetime
event approaches, Petri-net approaches, predicate transition systems
Process-oriented perspective operation of the system
¨
¥
Advantages
§
¦
development methodology and scheduling, results in development strategies (top-down, inside-out, ...), analysability
¨
¥
Disadvantages
§
¦
depends on whether a system will have this perspective
Development of Modelling Results
result of modelling: (c, d, (ρ, θ, ψ), g, w)
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
• representation concept c
• things d under consideration
• with restrictions for their applicability ρ
• with a rigidity or modality θ
• with a confidence ψ on their validity
• based on a common understanding of a group g
• within their world w or culture
Models and Solutions Imperfection
explicit modelling of the divergence from the real world
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
• incompleteness
• delay
• simple error
any model is imperfect due to
exceptional states events, time lags
incompleteness to to limitations of the language, consideration,
errors either based on real errors and exceptional states or based on
biases
Properties of Modelling
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
Monotonicity: any change leads to a refinement
Incrementality: any step is only based on new requirements or obligations and on the current specification
Finiteness: any quality criteria can be checked in finite time applying
a finite number of checks
Application domain consistency: corresponds to the requirements and the obligations of the application domain
Conservativeness: any model revision that cannot be reflected already in the current specification is entirely based on changes in
the requirements.
at least conservative and application domain consistent
any finite modelling process can be transformed into a process
that is application domain consistent
if the modelling process is application domain consistent then it can be
transformed into an incremental one if we can extract such area of change
in which consistency must be enforced
Orthogonal Subschemata
Address
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
Fond
6
?¼
Person ¾
Organization
Supplier
Customer
-
66
Product
6
Order
Personnel
Management
Production ¾
6
Budget
6
Budget
Department
6
Delivery
¾
¾
Billing
Reasoning on Properties
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
Subject of modelling depending on the application area (DSL), organisation units, scope, selected language, language level and expression style, categorisation, classification, terminology
Quality of the model depending on the modelling language, modelling transformations, formal treatment, formal derivation of properties, measurement
reason whether a model is appropriate or suitable, relative completeness (Frege) depending on the purpose of the model
depending on the quality of activities, goals, relevance, reusability,
maturity, understandability
Economy of modelling partially using standards, evaluating cognitive distance, evaluation value of the model, ROI
Separation of Concern: The Zachman
Framework
¨
¥
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Living in a world full of different models
§
¦
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
¨
¥
Coherence, calibration, mapping problems
§
¦
Model Suites: Collaboration
of Models
¨
¥
Handling abstraction
§
¦
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Explicit collaboration of models based on
• constructors
• mappings
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
• contracts among models
Dimensions of models based on the minimalisation of models and
constructors
Abstractions of models among mappings
Constructors for construction of new models
shuffle product, reduct , scope , integration
Theory extension for model context representation and context integration
Layers of Models for IS
¨
¥
Managing model suites by stratification
§
¦
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Collaboration specification layers
Infrastructure
Collaboration
Application domain
layer
Application environment
Collaboration policy, principles, acts
Requirements layer
System sketches, requirements, system decisions
Collaboration tasks, contracts, style, pattern
Business user layer
System view, parties, portfolio
Collaboration stories
Conceptual layer
Information system specification, context support
3C-C schemata,
informational
processes,
exchange frames
Logical layer
Information system - logical
view
Collaboration
system
Physical layer
IS programs
Collaboration programs
Content
Information
Concept
Topic
c
°B.
Thalheim
Deployment layer
...
...
supporting
Model Suite: Constituents
• set of models {M1 , ...., Mn } ,
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
• association or collaboration schema among the models,
Introduction
• application schemata for explicit maintenance and evolution of the
model suite, and
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
• controllers that maintain consistency or coherence of the model
suite,
• tracers for the establishment of the coherence.
Coherence describes a fixed relationship between the models in a model
suite.
¨
¥
only inductive languages with compositionality principle
§
¦
¨
¥
§concentration on discrete domains ¦
Model Suite: Languages
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
Model language L: signature S and a set of constructors C
ΣS,C well-formedness conditions
Model type TLS = (LS , ΣLS )
language of the model and
constraints ΣLS ∈ L(ΣWellFormed
)
S
Partial mappings Ri,j : LSi → LSj among LS1 , ...LSn
Model M: structM in LS
that obeys ΣLS ,
and set of constraints ΣM defined in the logics of this language.
Model Suite: Model Association and
Contracting
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
Collaboration contract among models
Collaboration
• Communication is used in a variety of facets as an act or instance
of transmitting or a process by which information is exchanged
between models through a common system.
• Coordination expresses the act or action of coordinating the harmonious functioning of models for effective results.
• Cooperation expresses the action of cooperating.
Collaboration style: supporting programs, data access pattern, style
of collaboration, coordination workflows
Collaboration pattern: supporting access and configuration, event
processing, synchronization, and parallel execution
Model Suite
Model suite type ST = (TLS1 , ..., TLS , ΣLS1 ,...,LSn )
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
model types TLSi defined on a set LSi , ..., LSn
ΣS1 ,...,Sn constraints
Model suite S
on a model suite type ST
models (M1 , ..., Mn ) of type TLSi
that obey ΣLS1 ,...,LSn
Contract on C:
• constraints ΣLS1 ∪ ... ∪ ΣLSn ∪ ΣLS1 ,...,LSn ,
• description of the enforcement mechanisms for any operation that
can be used for modification of one model, and
• set of consistent evolution transformations.
Model Suite: Synchronisation and
Coherence
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Explicit mapping among models
Mi
Mi
extract ei,jMi,to(j)
¾ load lj,i
Mi,f rom(j)
transform ti,j
-M
j,f rom(i)
transform tj,i
¾
Mj,to(i)
load li,j -
¾extract ej,i
Mj
Mj
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
Coordination
profile
modeli
ei,j
lj,i (tj,i (ej,i ))
ei,j
¾
tj,i (ej,i )
modelj
Model
Transformer
evolution-prone
completed to a model suite architecture
ti,j (ei,j )
¾
ej,i
- li,j (ti,j (ei,j ))
ej,i
Hierarchical Layered Model Suites
Typical example for models at the same abstraction level
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
auxiliary
database
?injected
DBS
Content
Information
Concept
Topic
c
°B.
Thalheim
insert
inject
modifiableinjected
micro
DBS
sensors,observations,
transaction data
auxiliary
database
?injected
DBS
(S c , ΣS c
O c , ΣO c )
(S p , ΣS p ,
Op , ΣOp )
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
auxiliary
database
?injected
-
insert
inject
auxiliary
database
?injected
DBS
DBS
(S r , ΣS r ,
Or , ΣOr )
modifiableinjected
meso
DBS
storage, capturing,
historical integration, leverage, archiving
-
insert
inject
analysis
DBS
analysis, mining,
exploration, hypotheses generation
explicit inheritance of underlying data
ownership principle
explicit explanation based on underlying data
agreed stratification of data and schemata
(S m , ΣS m ,
Om , ΣOm )
modifiable
-
-
injected
presentation
DBS
business
sheets,
appendix for publication,
web
presentation
Layers as a Building Block of
Multi-Layered Architectures
¨
¥
Generalising architectures
§
¦
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Architecture of a multi-layer model suite
Operators: Oi+1,1 , ...
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
Layer i+1
?
Data objects: ti+1,1 , ...
Layer i
Operators: Oi,1 , ...
Data objects: ti,1 , ...
“uses”
Oi+1,p (ti+1,q )
“realises”
-
6
Y
+ Layer language
Oi,r (ti,1 , ..., ti,k ), ..., Oi,s (ti,1 , ..., ti,m )
Dimensions of Data Modelling: Profile and
Portfolio, Abstraction and Extension, and
Quality of Data
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Data aggregation, abstraction and enrichment
6founded annotated aggregated macro-data
annotated aggregated macro-data
macro-data
micro-data
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
Application A
Application B
-
Data profile and portfolio
Gossip/raw/sensor/source data
Staged/cleansed data with data profiles
Consolidated and transformed data with hocks for data change capture
Integratable data ready for on-demand use in federations
Coherent data enhanced by services for in-line delivery in enterprise data farms
ª
Data quality improvement
Principles of Multi-Layered Modelling
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
Downward-dependency principle: The main data dependency structure is topdown. Objects at a higher level depend on objects at a lower level.
Upward-notification principle: Objects at a higher level act as subscribers to database changes at lower level. They may decide whether they eagerly or lazily
enforce observed changes at lower level. Objects at lower level report however
their changes.
Neighbor-communication principle: Objects may exchange data only at the same
layer with other objects. The neighborhood may also require that neighboring
databases should be synchronised.
Explicit association principle: The data exchange between databases is explicitly
documented and recorded. Whenever a database at a higher level perceives
data from a lower level then this exchange is logged.
Cycle elimination principle: Cyclic data exchange between layers is broken based
on the log information.
Layer naming principle: Data belong to their level and can be identified at their
level.
Structural and Functional Integration of
Models
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
Mappings and actual data mappings (synonyms, homonyms, model suite
hypernyms and hyponyms)
partiality of data mapping, abstractions
• Microdata as the starting point for collaboration
• Mesodata through abstractions, filtering, scoping, summarisation
• Macrodata for model injection
Handling missing, intentionally not available, not applicable, biased and low
quality data
Informorphism among different equivalent presentations
Function integration as generic functions depending on the model
Constructor integration depending on the data profile and on the task portfolio
Choices for Specification: Model Pattern
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
¨
¥
Depending on the style of development
§
¦
Structure-oriented pattern such as
• Compacting patterns
• Typing patterns
• Unfolding patterns
• Union patterns
IS rule pattern
separation variation state transition control virtual machine convenience
pattern
pattern
pattern
pattern
pattern pattern
¨
¥
Advantages
§
¦
efficient development with controlled refinement, repeatability, robustness, incrementality
¥
¨
Disadvantages
§
¦
restrictions in the development style, incrementality
Responsibility Pattern
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Regulations
requested by
Responsibility
type
responsible
¾
Party
type
6
Knowledge/strategic level
6
Temporal/tactical level
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
requested by
for
-
Responsibility
responsible
?
?
Action
Time
frame
-
Party
(0,n)
(0,1)
Evolution of Data
Comparison
operator
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
I
6
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
Duration
L
Observation
ª
?
Date
µ
-Measurement
R
*
-
R
State
¾
6
Real
state
- State type
Planned
state
- Planning
]
Method of
measurement
Planning
of state
Planning and Variations
Action
state
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
6
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
Relation to
time
performing
¾
-
Action
Á
]
Interested
party
?
Proposed
action
Location
Implemented
action
Summarising
¨
¥
... although there is not too much to summarise
§
¦
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
¨
¥
Instead of my own summary: The consternation summary at Modellierung 2009
§
¦
(1) ... but they do not know what they do ...
(2) Babylonian language confusion and muddle
(3) “It’s not a bug, it’s a feature”, de-facto-standards and lobbyists
(4) why I should cope with what was the state of art yesterday
(5) each day a new wheel, new buzzwords without any sense, and a new trend
(6) without consideration of the value of the model
(7) competition is a feature, inhomogeneity
(8) Laokoon forever
(9) dreams about a sound mathematical foundation
(10) take but don’t think - take it only without critics
Content
Information
Concept
Topic
c
°B.
Thalheim
(11) academia in the ivory tower without executable models
(12) where is the Ariadne thread through
¨
¥
Our goal: Overcome this situation! ¦
§
Thanks to
Conceptual
Modelling
Theory
Nov. 09, 2009
B. Thalheim
Introduction
Disconcertment
Dimensions
Modelling?
Language?
Modelling Act
Model
Principles
Theory of CM
Model Suites
Model Pattern
Content
Information
Concept
Topic
c
°B.
Thalheim
• the organisers (esp., Klaus-Dieter Schewe) of this workshop for their courage
to invite me to give a talk on a topic that never will get a complete solution,
• the (DB)2 , RADD and mainly the ID2 communities that gave me a deep insight
into their database developments and contributed to our schema library,
• to Roland Kaschek who developed the notation of the model on the basis of
Stachowiaks model theory and Pfänders theory of decisions
and finally got an insight into the development gap,
• the participants of the ‘Modellierung 2009” workshop with their intensive discussion of different facets, deficiencies, pitfalls and opportunities of a theory of
conceptual modelling
thus stimulating this research,
• the colleagues from the ICfTC providing me an insight into philosophical investigations, and
• finally everybody who is going now to contribute to a development of a theory
of conceptual modelling (Please let me know your results via email to
¥
¨
§thalheim@is.informatik.uni-kiel.de ¦
or let me join your research!)
Thank you!
thalheim@is.informatik.uni-kiel.de
Content
Information
Concept
Topic