Author Topic: Use Case Editor: Automatically Highlight classes  (Read 890 times)

AndyJ

  • EA User
  • **
  • Posts: 332
  • Karma: +5/-2
  • It's only a model
    • View Profile
Use Case Editor: Automatically Highlight classes
« on: February 22, 2010, 11:13:41 am »
Hi team,

I understand that you have extended functionality in the Use Case editor in release 8.0 to automatically highlight items that appear in the project glossary.

This is a very useful function, and I wonder if you could extend this behaviour to recognising classes in the domain model?

In our Use Cases, we keep the Basic and Alternative flows as simple as possible by referring to details in the domain model as follows:

The system displays a method to enter Search Criteria

Search Criteria would be defined as a class on the domain model, showing the expected sub-types of search criteria, each with its own distinct attributes etc.

This means that the Use Cases are very easy to read, without having to wade through details which are already defined in the domain.
Sun Tzu: "If you sit by the river long enough, eventually the body of MS Visio floats past."

salayande

  • EA User
  • **
  • Posts: 224
  • Karma: +0/-0
  • I love YaBB 1 Gold!
    • View Profile
Re: Use Case Editor: Automatically Highlight class
« Reply #1 on: February 23, 2010, 05:26:32 am »
I support this simple feature and more. Please, kindly raise a feature request.

Current best practice to domain concept models suggest the development of a domain taxonomy which guides the selection of the business concepts used in the domain model.

EA could parse the text of the business use cases and generate a candidate list of business concepts based on the tuple (subject, verb and Object).

An experience of "Sentence Diagraming" techniques (Check Arlow and Neustadt "Literate Modeling Techniques") suggests that above tuple may be applied to business use case statements to generate a candidate taxonomy. Use Google to search for a paper with the term "Taxonomic Class Modeling" (It wasa long time I read the article.).

My suggestions are already achieved in the areas of Ontology extraction from text corpus (Check OntoGen 2.0 for a prototype). I currently use the IBM Taxonomy tool to achieve the same purpose (Check IBM's web site for an evaluation copy)

I think that this approach would add a bit of intelligence to current approaches to class / concept modeling especially if this feature is backed by a thesaurus such as WordNet (Check Princeton University).

I hope this community will agree that we need a bit more intelligence in our tools.

kind regards

Segun