Semantic data model
A semantic data model (SDM, also conceptual schema ) is an abstract, formal description and representation of a section of the "perceived world" in a certain context (e.g. a project ) within the framework of data modeling . There are various modeling languages for the formulation of semantic data models, of which the entity-relationship model is the best known.
Different terms: For semantic models, other terms are used in practice, characterized by different methodological approaches or operational habits, for example: conceptual (data) model, conceptual database scheme, conceptual model, logical model, information structure.
purpose
- Most commonly it is during the design phases in projects developed software development - ultimately as a basis for the development / commissioning of a database .
- Area or company data model: Models are created for individual operational areas or the entire company, which should / can be used in individual projects as design samples or reference models. From this z. B. the specified designations, the textual descriptions or the designations with which the components in user media (such as screen or list displays) should / can be named (e.g. short / long form ...).
- Regardless of the objective 'model for databases', it is suitable to present any information context clearly and clearly, e.g. B. Roles, rights and participants in organizational studies, metamodels in method manuals, etc.
In terms of working methods , a semantic data model in data-oriented projects has the purpose of recording and displaying all the relevant technical aspects of the topic of 'data' as precisely and completely as possible and in a uniformly structured manner in early project phases that are not yet IT-technically determined. The following applies:
- There are only a few means of representation; these are easy to understand and use.
- These are i. d. Usually a graphic overview and textual descriptions for each term determined.
- All technically important aspects relating to the data identified as relevant are defined as precisely as possible.
- Which are there? What is meant by this? How do they belong together? What can, what must be? ...
- IT-technical implementation aspects do not yet play a role.
- Example: Form of data storage, access methods, only calculate information or also save it? ...
- The model is not a 'technical documentation'. Rather, professionally oriented project employees know / understand (after a brief introduction to the methodology) everything that is presented and described therein.
From the final point of view , the early creation of semantic models should lead to high data quality , which in turn determines the quality of the results of a project (a data storage system) and the quality of processes (efficiency in project work and in company processes). Data models also support corporate communication through their conceptual and structure-forming effect - in projects and in business operations, professionally and technically.
Components
The components of a semantic model depend on the respective modeling language. See also entity relationship model .
An example of essential work content when creating a semantic model:
- From the 'information terms' determined to be contextually relevant, units ( 'entities' ) are identified and - as ' entity types ' - named "ABOUT which" information must be processed and / or stored.
- Example (banks): Name, date of birth, account opening date, transfer amount, balance, interest rate, zip code are meaningfully combined to PERSON, ACCOUNT, BANK TRANSFER, LOCATION.
- It is determined which relationships (also contextually relevant) exist or can exist between these entities.
- Example: Each ACCOUNT belongs to 1 person; Conversely: each PERSON can have several accounts.
- The determined facts are described in text form and i. d. Usually shown graphically, e.g. B. by an ER diagram .
Form, content and terms of created data models can be very different and are e.g. B. depending on the following criteria (each with possible examples):
- Modeling methodology used: Models created according to UML look different than models created according to ERM .
- Purpose of creating the model: Database redesign requires more detailed descriptions than if (e.g. in a maintenance project) the data to be processed already exist.
- Modeling tools used: graphic form of relationships (rhombus or line ...), designation of terms (relationship, relation) differ depending on the tool.
- Project / company-specific rules: Level of detail of the modeling (e.g. for relationships); Specification "The data model is only created in the IT concept"
Modeling languages
The predominant language for describing semantic data models is the Entity-Relationship-Model (ER-Modell) conceived by Peter Chen in 1976 or one of its numerous extensions. In practice, simplified models such as Martin notation are also often used. In addition , the Unified Modeling Language (UML) is used , especially for object-oriented modeling .
research
The International Conference on Conceptual Modeling (ER, formerly International Conference on the Entity Relationship Approach ) has been held since 1979 (annually from 1985 ).
See also
literature
- Graeme Simsion: Data Modeling. Theory and Practice . Technics Publications, 2007, ISBN 978-0-9771400-1-5