hiding the parent class name in embedded classes

[David E Foster]

Hi there. This is my first post. I have just started using EA for domain modelling and so far so good however...

When I embed a class (e.g. Person) in a generalising class (e.g. Party) the displayed name is changed from 'Person' to 'Party:Person'. Is there anyway of getting the displayed name to not change?

Thanks

David

[Geert Bellekens]

David, you can't, but why in deus name would you want to embed Person in Party  , especially if Person is a subclass of Party.

Geert

[Paolo F Cantoni]

Hi David,

To echo Geert, there was a discussion (within the last month) on exactly this topic.  If you search for it you'll see the previous viewpoints.

If you find it, perhaps you could post a link here.

Paolo

[RoyC]

And, users of the forum seem to have much more success in finding earlier posts using the left hand Search button at the top of the page, rather than the right hand Search field.

[Paolo F Cantoni]

And, users of the forum seem to have much more success in finding earlier posts using the left hand Search button at the top of the page, rather than the right hand Search field.

Verily, thine lips hath said it...

(Take the hint and make it clearer)

Paolo

[skiwi]

http://www.sparxsystems.com/cgi-bin/yabb/YaBB.cgi?num=1280280983

Now how about we get an improved version/configuration of the BB that by default has a better search top right, and better defaults (eg for 6 months) for for search top right button, and allows more than 100 results to be returned

[David E Foster]

Example 2 here http://www.tdan.com/view-articles/9219 is what I am trying to achieve. I thought that this would be achieved by 'embedding' in EA terms (I assume that have I got that wrong?).

[sargasso]

* Sound of awakening dinosaur

David,

David Hay is a very very clever man who, some time before the last ice age, provided a vary convincing argument that UML and data modelling are not the same thing. In short, that a bicycle is not a fish.

The article you cite seems to be arguing that a bicycle could be a fish if it had gills and fins.  It is not, IMNSVHO, a great deal of use.  Fishes are fishes and bicycles are, well, a subclass of two wheeled human propelled conveyances.

The diagram is illustrative only in his argument for the need for a data domain modelling technique suitable for non-technical understanding of a particular domain.  It is not an "accepted" use of EITHER static data structure modelling, nor of a UML information domain.  While he may be correct in specifying a requirement for such a modelling technique, the article does not specify a way that the current status of UML can achieve such a model.

The picture is not the model.
The model is not the system (nor the data).
A fish is not a bicycle.

.... going back to sleep now...
bruce

(b.t.w. Did you know that Spectrum (or alternatively "The Incredible Murtceps" was possibly one of the greatest bands that ever graced this planet   )

[Geert Bellekens]

David,

I'm afraid you indeed "have it wrong" and that fact alone is probably a reason why not to use the notation suggested by the author of that article.

Onto the subject, the author is actually talking about regular Generalization relations between Party and Person, but he is just drawing them onto each other on the diagram. Their location on the diagram doesn't say anything about their location in the model.

If you really want to achieve this with EA then you just need to drag the boxes on the diagram on top of each other, without touching the location of the classes in the model.

Geert

[Paolo F Cantoni]

[size=18]...[/size]
If you really want to achieve this with EA then you just need to drag the boxes on the diagram on top of each other, without touching the location of the classes in the model.
[size=18]...[/size]

But further to what Geert says, and to take issue (even with David Hay), you mustn't do this on a Composite Structure Diagram...

"A composite structure diagram is a larger rectangle, with its components contained as rectangles within it".  It displays composition, NOT generalization.  It's one of the few UML diagrams where location of items within other items is meaningful - and it's meaning is NOT sub-typing.

Paolo

[David E Foster]

Thank you Geert,

Turning off 'support for composite objects' had the desired effect. I am presenting a domain diagram that is for presentation to non-technical end-users and so ease-of-understanding is my paramount concern over technical notation correctness. I appreciate your comments though.