Again: How to model delegates in object diagrams?

[Hans]

Sadly, I didn't receive any answer on that.
And, in view of the property/field discussion in here, I did not expect that.
When I decided to use EA (and by this UML) this was although I knew of the
severe shortcomings of the narrow-minded Booch-IBM-Java-orientated conception
of UML 2.0.
My stimulus was and is that UML is a means for us to standardize and
extend.
Of course, I do know that delegates (callbacks) are out of scope (as by Booch)
as are properties, but that should not hinder EA to discuss and propose standards
and best practices for all that stuff beyond: We simply want to build more than
Lada motors!

To the topic: There are some CASE-tools which do supply delegate and property
items besides attributes and operations. That could be one solution. A more
conservative one would be to introduce stereotypes for delegates (and properties).
But then the problem is to the nature of a delegate: Is it an attribute?
And what about the subscriber? How do I model that?

I think, it is to us and Sparx to find the proper answers. Sparx might perhaps
devote a forum to these points and extract the wisdom into its own idioms.
But to sit away or ignore these problems simply doesn't count!

[sargasso]

Try searching these forums for delegates - for my name sometime earlier last year.  I had a similar problem with C# delegates and if I recall correctly there was an answer.

Its not a huge item of interest.

Bruce

[Hans]

Thank you, Bruce!
That solved my question.

Regards,
Hans

[Hans]

Hi Bruce,

one question: what do you mean by

"Its not a huge item of interest. "?

Of course, I can take the GoF subscriber pattern and duplicate  everything in there, sit back and say to myself: Great how that looks, and how beautifully all these nuts and bolts fit!

But actually, it stinks, when you've got better options at hand - if they are Booch-blessed or not!

I think you're of the same mind: It should matter that we get our tasks done with the same UML whether it's in the Unix world, OS/2, or whatever: it's that the means we got at hand in there are different, and a UML shall have to account for that.

Regards,
Detlef

[sargasso]

Oh, I was just referring to

Sadly, I didn't receive any answer on that.

Once I had sorted out the delegate structuring to my satisfaction whenever that was, I find that I have no need to model it explicitly anymore.

BTW if you found that topic, could you post the link here in case someone else needs it - I have no recollection of what I called the topic.

bruce

[benc]

This might have been it

http://www.sparxsystems.com.au/cgi-bin/yabb/YaBB.pl?board=general;action=display;num=1082338556;start=6#6

I believe the minor code template bug I referred to has been long since fixed.

Regards,
Ben

[Hans]

Yep, that's it!