Information Flows - Multiplicity

[Viking]

Hello,

I am trying to create something like a metamodel for the relationship "Information Flows".

I want to show (e.g. in form of multiplicity) that it is allowed to attach zero to many Information Item Elements to the relationship.

Is there a possibility?

Many thanks in advance, V.

[qwerty]

That leaves too much open questions. What are the conditions? How should it be shown? Usually you would run a post-check in an add-in to tell that the connector creation failed because of... Telling that in advance seems pretty much undoable (in EA with acceptable effort).

q.

[Viking]

That leaves too much open questions. What are the conditions? How should it be shown? Usually you would run a post-check in an add-in to tell that the connector creation failed because of... Telling that in advance seems pretty much undoable (in EA with acceptable effort).
q.

You are right. I should have spent more time on the description of my issue.

As a picture is worth a thousand words:



V.

Rem.: the picture is not shown on every browser. If not, click here: http://imageshack.com/a/img921/5773/gXUSaX.png

[qwerty]

You seem to be mixing it with an association class (your m/n labels are not correct in that case). I don't think that's a good idea.

q.

[Viking]

You seem to be mixing it with an association class (your m/n labels are not correct in that case). I don't think that's a good idea.
q.

The picture shows only the idea what I am looking for. There is nothing "decided" or "implemented".

[qwerty]

Sure. But I don't get the rationale behind it. Would you like to show the information items in a different way? Then why does the lower picture use a class named Classifier?

q.

[Viking]

Sure. But I don't get the rationale behind it. Would you like to show the information items in a different way? Then why does the lower picture use a class named Classifier? q.

"Classifier" is wrong. It should be "something" representing Class 3, 4, etc. I am sorry.

[qwerty]

Hmm. The only thing that comes to mind it to replace the label holding the list of transported items with something else in a connector shape script. As I see it you would need to add-in to do calculations (in order to show the m and n). Shape script alone can't do that. Even if you get that add-in to work (at least it's possible) you might struggle with finding a fashionable format. That association class-like linked element is a tough one (if doable at all).

q.

[Viking]

Hmm. The only thing that comes to mind it to replace the label holding the list of transported items with something else in a connector shape script. As I see it you would need to add-in to do calculations (in order to show the m and n). Shape script alone can't do that. Even if you get that add-in to work (at least it's possible) you might struggle with finding a fashionable format. That association class-like linked element is a tough one (if doable at all). q.

Many thanks, qwerty.

Your suggestion sounds very difficult.

I also thought about the Association Class. Maybe this is a way. But I cannot add multiplicities.

[qwerty]

Well, an association class has a defined semantics. But you could create a stereotype for an AC which makes them holding the conveyed items as attributes. This AC could have a shape script showing the number of already added information items (which again would need an add-in for the calculation). In any case, this would be a very specific solution only applicable in your domain. And thus people need to learn about it. I don't know the rationale behind your question. Sometimes there's a demand which calls for individual solutions. However, I'd try to stick to standards as far as possible.

You should probably think of hiring a consultant to help you with that task and discussing it in detail. (Not me. Ain't got time.)

q.

[Viking]

Well, an association class has a defined semantics. But you could create a stereotype for an AC which makes them holding the conveyed items as attributes. This AC could have a shape script showing the number of already added information items (which again would need an add-in for the calculation). In any case, this would be a very specific solution only applicable in your domain. And thus people need to learn about it. I don't know the rationale behind your question. Sometimes there's a demand which calls for individual solutions. However, I'd try to stick to standards as far as possible. You should probably think of hiring a consultant to help you with that task and discussing it in detail. (Not me. Ain't got time.). q.

Thank you, Q.

This sounds even more complicated.

I thought about the meta model of UML. That should show this view(point) as classes. But I cannot find it in the meta model.

Do I think wrong?

[qwerty]

That leaves me with a question mark. UML allows to extend its provided metaclasses via stereotypes in profiles. Of course these stereotypes are not present but must be defined by people using UML. So not sure what you were looking for.

q.

[Viking]

That leaves me with a question mark. UML allows to extend its provided metaclasses via stereotypes in profiles. Of course these stereotypes are not present but must be defined by people using UML. So not sure what you were looking for. q.

That leaves me also with a question mark. My fault!