Model

Models are created in the areas of application in the modeling software, Innovator, normally in the form of domain-specific conventions of graphic elements.

Properties and What Models are Used for

A model is a simple map of reality. According to Herbert Stachowiak, there are three characteristics of a model:

  • Illustration

    A model is always a map of something, a representation of a natural or artificial original which can itself be modeled.

  • Summary

    A model does not contain all the attributes of the original object being modeled; it only contains attributes which are relevant to the person who creates the model and those who use it.

  • Pragmatism

    Pragmatism essentially means to concentrate on the business side of things. A model is related to the original and not itself. Assignment is qualified by asking the questions "For whom?", "Why" and "For what purpose"? A model is used by the person creating the model and those using the model within a specific time frame and for a specific purpose for an original. The model is interpreted based on this.

(Source: Herbert Stachowiak: Allgemeine Modelltheorie (General Model Theory). Vienna 1973, ISBN 3-211-81106-0)

Within IT, models are used for mapping a section of reality to help complete a task using information processing. Models of the same type are called domain models. Included in these are e.g. models for software to be created, for both architecture (architecture model) and code (in the form of e.g. program flow plan diagrams), data models for describing structures of data to be processed from the business/technical side of things or from a technical data management view.

Models can also be used as a template when designing an IT system; these are then known as system models. These include reference models which can be used as a design template. Reference models are used for e.g. designing concrete computer systems, network protocols, data warehouse systems and portals.

As well as these models which manifest themselves in databases, there are also models for planning, controlling and organizing. Typical objects to be modeled include the flow structure of business processes mapped in a business process model and how an operational organization is structured mapped in an org chart.

(Source: Manfred Broy, Ralf Steinbrüggen: Modellbildung in der Informatik (Modeling in IT). Springer, Berlin Heidelberg 2004, ISBN 3-540-44292-8)

Software projects only contain abstract models. Three main aims of abstract models in software development can be identified:

  • Ensure that requirements are met and you have a sound system design by testing the model

  • Support the solution orientation system planning. Code can be generated directly from the model in an ideal case

  • Make communication between users, developers, analysts, testers, managers and anyone else involved in the software project easier

These aims are met through abstraction in the real world. This is how e.g. an abstraction portrays various levels of a system, another abstraction portrays the interaction between users and the system or interactions between various systems.

Different abstractions are necessary as the needs of the different actors in a model vary. A user expects a model to show them future interactions, whereas a developer is more interested in objects which need to be designed. A tester, on the other hand, expects a model to show them interactions between objects so they can derive their test cases from them.

A model should always provide as much information as possible for each different group using the model. The target group the model is aimed at should not be the only ones to get a better and deeper understanding of the system being examined; the person who created the model can also gain new perspectives and recognize new relationships which they didn't see before when modeling (this is particularly helpful for business process engineering).

(Source: Sang-Il Kim: Modellierung von Workflow-Prozessen in der Radiologie (Modeling Workflow Processes in Radiology). Universität Hamburg, Diplomarbeit 2003)

The concept of a notation is defined by their abstract syntax and semantic, i.e. the language elements and their syntactical assignment. They provide an interpretation and form the formal basis for the modeling language. The abstract syntax and semantic form the actual language description.

The concrete syntax then sets textual and graphic language elements based on the abstract concept. Graphic display forms can dominate the concrete syntax. The notation sets the graphic representation of the language concept but links to individual notation elements are only permissible based on the abstract syntax and semantic.

Elements and their relationships make up a model. These are graphically shown using specific nodes and edges. Diagrams and tables are collections of elements and relationships which belong together.

The respective language, the notation, controls which elements and relationships need to or can be part of a model. Modeling methods specify how notations should be implemented. One of the biggest difficulties when modeling and the error which has the biggest consequences is forced abstraction of objects from the real world, i.e. their forced splitting and simplifying into aspects of requirements for the model.

Elements and relationships have general and specific tasks in the model. This is why they have a certain type which groups them and defines them into basic properties. Element type and stereotype are terms which express this grouping.

A method is a process which can be planned and that helps when solving theoretical and practical problems. Good methods can be recognized by the fact that you can immediately see how they refer to the problem, that they increase the quality of how the problem is solved and that it is obvious which prerequisites need to be used for them.

Modeling languages define semantic, syntax and notation of elements at a meta level which can be used by the elements. It makes a lot of sense to add typical extensions to concrete application fields.

A modeling method requires one or more languages to reproduce the necessary views in a way that can be understood by those viewing them.

  • Technical terms and concepts should be set using domain-specific languages (DSL)

  • Languages should support systematic creation of consistent documents

  • Languages should be able to prepare software design, i.e. have a formal syntax and semantic

Support of international standards and modeling conventions enables processes which can be divided into jobs that need to be carried out; it also enables their results to be carried out seamlessly and be comparatively easy for all those using it to understand and use.

Standardized modeling methods have the following advantages:

  • High level of transparency

  • Comparable project results

  • Common basis for discussions for those working in groups in a project

  • Good basis of information (internet, books, application fields)

  • Extensive independent experts

If various modeling methods are combined from the view of a process role into a desktop and transformed with extensive similar functions and common tools, this then means that it is easier to achieve a hight level of productivity and systematic daily use. These, in turn, are prerequisites so that models can be kept up-to-date and offer the support needed.

Modeling tools portray structures and flows systematically and, in this way, document various views of the relevant sections from reality.

Modeling tools should be powerful, diverse and as easy to use as possible. They need to ensure traceability which comes with modeling linked abstractions; this is done by structuring the model itself into transparent structures, controlling the namespace and permitting descriptions of model elements using texts, relationships and strictly-structured properties.

Modeling tools normally focus on one modeling method and improve their functions for this specific application. Those involved in a process who have various tasks and for the project in general, it is normally necessary to implement multiple tools with various different uses and solve numerous interface problems.

Integrated modeling software takes a different approach; it provides a combination of methods with uniform user support and facilitates embedding in system landscapes with flexible internal and external interfaces.

Modeling tools are used for actual modeling, i.e. implementing objects from reality into abstractions of a statistical and/or dynamic model; they are also used for model documentation and specifying and visualizing model components and the applications they cause.