Circle notation for interfaces

[thomaskilian]

3) You're probably right. And EA offers the exposed iface icon only in the components toolbox.

[Geert Bellekens]

1) Just to top the [pedantry].. In my spec (just dowloaded the 2.1.2 no changebasrs 2-11-2007) the topic is numbered 7.3.24 (I don't even find a 7.15.3). Anyway, I think you better quote the whole paragraph there:

The interface realization dependency from a classifier to an interface is shown by representing the interface by a circle or
ball, labeled with the name of the interface, attached by a solid line to the classifier that realizes this interface (see Figure
7.55).

The important part here is "representing the interface by a circle or ball". So we are representing the actual interface. (or we should be)

3) Indeed that is exaclty what I use it for, but I don't think it is really relevant to the discussion. Whether it is on a component or a class doesn't really matter. It still feels a bit akward.

[sargasso]

(just dowloaded the 2.1.2 no changebasrs 2-11-2007) the topic is numbered 7.3.24 (I don't even find a 7.15.3).

Geert, absolutely correct!  With all the versions around now, I guess I was just looking at the wrong one.

What I'm (still pedantically) trying to say is:

"The <<connection>> between the <<bicycle_rider>> and the <<bicycle>> is shown by representing the <<fish>> by a <<squiggle>>".

(Which makes about as much sense!!) However the point I am trying to illuminate, is that the <<squiggle>> represents the connection not the <<fish>>.)  But, heh! I am the last person to suggest that a committee can speak English.

(Sigh) The 1.3 spec, as self-contradictory as it was, was at least "readable".(/Sigh)

bruce

[sargasso]

BUT! To get back to the core issue.

If (and only if), I am correct in that the lollipop notation represesents the realisation and not the actual interface, then,  there is no reason why you cannot have n classes adorned with the same realisation and an element describing the actual interface on the same diagram.

We could have lots of different, say "textboxes", on a diagram that all realise "_I_am_a_textbox".  A red one, a blue one, perhaps even a duggite one.  There is no reason that I can see that one cannot include an element  ("round" or square) on that diagram that describes in detail the actuality of the interface.

So Paolo, I concur, in fact I disagree!

bruce

[sargasso]

3) Indeed that is exaclty what I use it for, but I don't think it is really relevant to the discussion. Whether it is on a component or a class doesn't really matter. It still feels a bit akward.

There is actually a big difference between exposure and realisation between the class and the component.  I have writ much elsewhere, but between you and I, I still aint too sure what they were gettin at.

bruce