The word conceptual implies an abstraction away from the messy complications of reality. It suggests a pleasing and common-sense fit with the way something is generally understood. That is the case here, and differentiates the process of conceptual design from logical design which we will discuss later.
Recall that the ANSI-SPARC model has three levels, external, conceptual and internal. The process of conceptual design is concerned with establishing the data model at the middle level which is also known as the community view. The reason for this additional name is that this level must capture all the user requirements in terms of data and structure in a single, well-structured way. How do you check that the community view is accurate? You check with the community, of course. This means that your conceptual model may include features such as : relationships that are not possible to implement directly with relational technology if that is how the users understand the data. The conceptual model, therefore is all about communication: it is your synthesis of the information you have collected from users, and it must correspond with their understanding of the subject area.
As well as : relationships, we have also mentioned one or two other examples of structures that require further work before they can be implemented with relational technology. One of them is the generalisation/specialisation relationship that may exist between entities. This relationship might be explicit in a conceptual model, but might be implemented in different ways by a set of database tables.
Connolly and Begg suggest the following process for conceptual design:
- Identify entity types
- Identify relationship types
- Identify and associate attributes with entity or relationship types
- Determine attribute domains
- Determine candidate, primary and alternate key attributes
- Consider using enhanced modelling concepts
- Check model for redundancy
- Validate conceptual model against user transactions
- Review conceptual data model with user
All of these steps should now be familiar. If not check back to the relevant section in these notes, or read section 16.3 of Connolly and Begg.