Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.

Topics - qwerty

Pages: 1 ... 8 9 [10] 11 12 13
Bugs and Issues / Importing a whole model from an XMI
« on: February 25, 2012, 02:30:13 am »
Maybe it's just me being too brain damaged. However, I just made the following observation: From a (not too small) model which I copied from somewhere I imported the root XMI which contains the whole model. Right after the import I ran a baseline compare and - oh wonder! - got a lot of differences. Not only the style of many connectors was indicated to be different (why? a fresh import of the very same file). There were also complete packages missing!!!

So I re-imported the XMI into that model - now not from scratch but with most (!) of the elements/packages already existing. This import also imported the missing packages! (The style of a couple of connectors still is indicated to be different, though.)

So what is this? Normal behavior? Strange earth waves? Has anyone else encountered this?


Bugs and Issues / Actor "lifeline" in sequence diagram
« on: February 23, 2012, 03:42:26 am »
I noticed that when using an Actor from the toolbox the diagram does not show the classifier but only the object name (and there without leading colon). Is that only for my installation or could someone confirm that behavior? If so I'll submit a bug report.


Bugs and Issues / Information Flow Realized in messages
« on: February 23, 2012, 04:30:41 am »
To show information items flowing at an association you need to combine a connector/association with information flow(s) via Information Flows Realized in the advanced context options. That will give you a filled triangle showing the item(s).

Now in sequence diagrams you seem to be able to do the same. The Advanced/Information Items Realized is available. You can also attach information items and this results in a dashed connector free flowing near the message. But each time you try to enable the flow again via the context it does not change anything.

I thought it would be a nice feature to show an information item flowing as message instead of an operation of the receiver. However, I haven't seen anything the like in Superstructures. So the EA implementation seems to be a bug.



Bugs and Issues / Moving association/information flow buggy
« on: February 10, 2012, 10:13:08 pm »
I just submitted a bug related to moving a connector with information flows realized. Here are the steps:

- create two classes A and B
- draw an information flow from A to B
- choose Add and create a class C and close the dialogues
- draw an association between A and B
- for the association choose Advanced/Information Flows Realized
(note that its quite annoying that both connectors overlay each other and you do not get the context menu for the association immediately. Instead often you must shift one connector to reach the other which is very user unfriendly)
- now check the C class
- create a class D
- move the source end of the information flow/association from A to D
- note that the displayed connector shows connected from D to B
- select class A
-> in the Relationships window you see that  the information flow has not been moved!

As a work around I hide the association, move the information flow too and unhide the association. I lost quite some information flows after deleting the A classes until I noticed the source of the problem.

This handling is by far not optimal. I think there could be much better solutions to handle realized information flows.

I guess not too many use this feature as it's really hard to create and handle these associations. But to those which are using it: what do you think? Are you happy with the current handling?


Bugs and Issues / Opening Structured Scenario
« on: May 06, 2011, 10:34:40 pm »
When you open the Scenarios tab of a Use Case it seems that EA chooses to open the Description or Structured Specification at random. At least I could not find any rule which one is opened at first.

I thought it would make sense to open Description if it's not empty or Structured Specification vice versa (probably giving the later priority if both have data).

Should I report this as bug or do I miss something?


Bugs and Issues / Missing Note (Text, etc.) elements in browser
« on: March 31, 2011, 01:50:58 am »
As you know there are a couple of elements which do not appear in the project browser. These are at least Notes, Texts and Boundaries. These explicitly belong to a single diagram (well, more or less as you can re-use them in other diagrams too). However, once you create those elements they lurk around in the model SOMEWHERE. Usually the place is the package where the parent diagram had been located. When you move the diagram, those children stay in the old package. And you don't have a chance to move them along with the diagram. And that is bad.

When I create a documentation I sometimes like to render notes as text along with the diagram for readability reasons. But when the notes are not "nearby" the diagram it will confuse the reader.

Now my question: why not having these elements as siblings of diagrams in the project browser? I don't think it is correct to hide just some elements.

Any opinions?


Uml Process / Definition of Decomposition Icon
« on: December 06, 2015, 10:40:34 am »
Does anyone know where the Decomposition Icon in UML 2.5 is defined? I found only one reference in Fig. 14.8 on p. 320.


Uml Process / Can instances head lifelines in sequence diagrams?
« on: November 08, 2015, 07:42:54 am »
I stumbled over this interesting question on LinkedIn:

Now looking at EA there has always been some (confusing) mix of instances and lifelines. But if I interpret the current UML spec correctly lifelines and instances are different concepts. And thus instances must not appear in SDs.

Uml Process / Constraints in use cases
« on: September 23, 2015, 06:56:30 pm »
I needed to show pre- and post-conditions for use cases on a diagram and thought it would be neat to use the Link to Element Feature. But this raised two issues:
  • The notes don't seem to show which kind (pre/post/invariant/OCL) the selected constraints are. Any way to show them except adding an additional manual note?
  • Right after the constraint text EA renders a colon. What is the rationale for this?


Uml Process / Association ownership
« on: May 18, 2015, 09:43:32 pm »
In SS 2.5 (ptc/2013-09-05) they introduced a dot on the association to show ownership (p. 209). By accident I got one association to show that dot. As for that time I did not know its meaning I got rid of it be deleting the connector and creating it new. Now, after reading about the meaning, I would like to know how to create that dot.

Uml Process / Use of XMI 1.1
« on: April 13, 2013, 02:56:51 am »
I'm currently writing a compiler that creates a XMI ready for import by EA. The default is XMI 1.1. It looks pretty much straight forward so the XMI is imported by EA with no issues. However, I wonder how long XMI 1.1 will stay in place. At OMG there does not seem to be any legacy document left over for 1.1, they start at 2.1. Does that mean that XMI 1.1 is obsolete and I should move to at least 2.1? Or maybe someone can point me to the XMI 1.1 specification?


Uml Process / Exposed functionality from interfaces at ports
« on: April 30, 2013, 11:45:59 pm »
I have a rather strange issue. Or simply my memory is deceiving me. I tried this with v10 and 9.3: I have an interface (say IF) which provides some methods. And a class CLS with a port which itself contains a provided interface (say PIF). PIF is not only classified but has a Realize relation to IF. Now when I put CLS onto a sequence diagram and send a message to it, the drop down does not show the methods available from IF. Only if I show PIF within its port attached to CLS it is possible to utilize the methods from IF. But that rather clutters the sequence diagram. I'm pretty sure that I've seen the IF methods even at the CLS instance in a former EA version. Seemed logically to me as the CLS should have all the methods available it provides through its various exposed interfaces.

Another peculiarity which hit me here: it is not enough to classify PIF with IF in the Exposed Interface dialog. You must draw a Realization from PIF to IF in order to see IF's methods at PIF. So why create the classification if it does not provide the methods? Just for the name?

I'll try that with a former EA version (if I find the download link). However, it would be nice to know what's happening here.


Uml Process / Show includes/extends in an activity diagram
« on: April 29, 2013, 09:15:30 pm »
I got stuck with this one so maybe someone has a nice solution: In a use case "UC A" a single step in a scenario might read "The actor X performs the use case 'UC B'.". In a UC diagram you see either a <<includes>> or <<extends>> relation between both UCs. But how would one show that in the activity diagram for "UC A"? Assuming you have only a single activity in use case "UC B" and named this "Activity for UC B" (which from the "UC B" perspective looks like an over-definition) you could use that as a link element appearing in "UC A"s activity diagram. But probably there's another way to show that?


Uml Process / Interface? Interface!
« on: March 20, 2013, 05:04:10 am »
I wonder what the difference might be:
  • A class with stereotype <<interface>>
  • A metatype interface (3rd entry in the class toolbox)
  • A metatype interface with stereotype <<interface>> (which you get if you use Element/Advanced/Change Type from class to interface if the class had the <<interface>> stereotype) [see below]

Two additional observations:

  • I also noticed that right after changing a stereotyped interface class to an interface metatype it will show with the stereotype in the browser until you select it some times. It appears that now the stereotype is removed and it's just the interface metatype left over.
  • I also turned on Show Realized Interfaces and (like some other guy recently) noticed that the lollipop can't be moved - what the heck? When using auto-layout it's additionally shifted out of sight to the left!


[edit] Additionally, you may create a metatype Interface with stereotype <<interface>> (somehow that was created via add new element inside the select type dialogue for a parameter?!). So, what is this then?

Uml Process / UseCase Relation in use case diagrams
« on: February 08, 2012, 08:52:12 pm »
I wonder where from the UseCase Relation stems which is defined in the context of use case diagrams.

We have some kind of philosophic discussion about it and it is argued that a "use" in classical sense is connectd with Dependency. But this UseCase Relation is obviously a meta class of association (due to the fact that it is a solid line). Now I browsed Superstructures 2.3 and 1.5 but non seem to mention such a UseCase Relation.

Any light out there?


Pages: 1 ... 8 9 [10] 11 12 13