|
|
Rank: Advanced Member
Joined: 6/19/2007
Posts: 44
Location: USA
|
Hi,
First of all, I have to say that after reevaluating several other tools for software architecture with .Net, Umodel 2008 still scores way above the others and I hope that it continues to grow.
Ok,
1) Is there anyway to show the Internal Classes with delegates, etc. within a Component itself like this:
http://uml.netbeans.org/qa/nb55/ComponentDiagram/Plan_CPD_Usability_Images/Plan_CPD_Usability_html_7b821136.png
This would be a great feature to have. Currently I'm using ports to show internal classes which may not be elegant.
2) Also, it would be nice to be able to navigate directly to Types/Jump to code from Lifelines in a Sequence Diagram.
Thanks,
cp
|
|
Rank: Advanced Member
Joined: 9/28/2005
Posts: 371
Location: Vienna
|
Hello cp!
Thank you for your praise and of course we will try to extend UModel continously (IMHO the last release 2008r2 again introduced some interesting and cool features).
1) Do you mean as attached below?
2) When working in the integrations (VS.NET / Eclipse) this should already be possible!?
E.g. when you right-click a lifeline of a class, you can use "Jump to code" to directly come to the defintion of the class.
Same for call-messages, where you can directly jump the operation definition in source code.
cor attached the following image(s):
|
|
Rank: Advanced Member
Joined: 6/19/2007
Posts: 44
Location: USA
|
Hi cor,
1) Thanks for the diagram. However, could Classes be shown inside the Component instead of the sub-components so that the it can be visualized as to which Classes within a Component physically realize the interfaces provided by that Component?
2) You are right about the ability to "Jump to code"..I guess I meant "Select Type in Model Tree"
Regards,
cp
|
|
Rank: Advanced Member
Joined: 9/28/2005
Posts: 371
Location: Vienna
|
Hello cp!
1) Of course you can also drag&drop (=> show) classes inside a component in a diagram! Or do I misunderstand something!?
2) for (call-) messages, it is currently possible to jump to
a) the code file/position where the operation is implemented
b) the message in the modeltree
c) the operation in the modeltree
For representants of a LifeLine a) and b) is also possible - c) is currently missing and we will add this ASAP.
Thank you for pointing this out!
|
|
Rank: Advanced Member
Joined: 6/19/2007
Posts: 44
Location: USA
|
Hi cor,
1) Drag&drop of Classes inside a Component in a diagram is not possible currently, where drag&drop means that the Classes get contained within the Component and move with the Component. The drag&drop of Components inside another Component seems to work fine.
As an added suggestion, maybe you can also give an option to create ComponentRealization relation for a Class when the Class is dropped inside a Component in a diagram; this is currently possible in the Model Tree.
2) This is clear, thanks.
Regards,
cp
|
|
Rank: Advanced Member
Joined: 9/28/2005
Posts: 371
Location: Vienna
|
Hello cp!
Drag&drop of Classes inside a Component in a diagram is not possible currently, where drag&drop means that the Classes get contained within the Component and move with the Component.
Yes, but this also meens that one component is nested in the other - something which might not always be wanted for classes or interfaces.
The drag&drop of Components inside another Component seems to work fine
Currently you can select both (component and classes inside) or simply use a selection rectangle to select the component and everything inside to be able to move everything together.
Anyway - we look into some improvements in that direction!
As an added suggestion, maybe you can also give an option to create ComponentRealization relation for a Class when the Class is dropped inside a Component in a diagram; this is currently possible in the Model Tree.
Right, but I can imagine that some want to show one class in different components on a higher modelling level, where automatically created ComponentRealizations can cause some problems / lack of understanding.
Do you already use our new feature of automatically generated ComponentRealizations?
|
|
Rank: Advanced Member
Joined: 6/19/2007
Posts: 44
Location: USA
|
Hi cor,
As you suggested, it seems that the best way to show the internal workings of a Component right now is to simply place Classes, etc. in its frame and select+move them around.
Right, but I can imagine that some want to show one class in different components on a higher modelling level, where automatically created ComponentRealizations can cause some problems / lack of understanding.
I might not be understanding this, but currently UModel allows to have only 1 ComponentRealization per Class, and it also allows to drag+drop Class into a Component in the Model Tree which automatically creates a new ComponentRealization, and so I was simply suggesting to have this ability in the diagram as well since a Class can be contained in only one Component currently. And yes, I've been using a lot of your automatic features within VS 2005 plugin, and they are really efficient for the most part...however it does raise many unhandled exceptions now and then especially when renaming diagrams or, to put it vaguely, if you handle it a bit roughly.
Anyway just FYI for future reference, here is a UML 2.0 diagram which shows Classes with special UML stereotypes to describe their roles. Thanks for the responses.
Regards,
cp
cp attached the following image(s):
|
|
Rank: Advanced Member
Joined: 9/28/2005
Posts: 371
Location: Vienna
|
Hi cp!
...only 1 ComponentRealization per Class...
That's what I meant in my previous post. If somebody wants e.g. to show that one class is used in a client component and in a server component and wants to show this on a "higher modelling level" and not in the physical one, where the class is implemented in a common library, such automatic componentrealizations can lead to some problems.
...however it does raise many unhandled exceptions now...
Not sure if you speak about bugs but if so, please contact our support team that we can look into this ASAP.
|
|
|
guest |
