Messages

1

General Board / Re: extend selections

« on: March 17, 2004, 08:19:01 am »

yes, that's all I found out. But it would be handy to add a whole bunch of new elements to an existing selection (like e.g. in powerpoint with shift + drag mouse (draw boundary) ).

2

General Board / extend selections

« on: March 17, 2004, 07:40:28 am »

a rather simple question....

how do I extend a selection of elements? Control + left mouse only adds one element per click and the boundary box is limited to elements that are all in the same region. Is there something like shift + boundary box to add a new selection of elements to an existing one? Whould be very helpful (can't imagine it doesn't exist but I haven't found it yet.    )

Thanks,

Ferdinand

3

General Board / Scenarios as instances of use cases

« on: March 01, 2004, 10:48:50 am »

Is it possible to create a graphical representation of a scenario as an instance of a use case? If I insert a use case via "as instance of element(object)" I only get the use case itself. (The same operation on actors gives me an instance of this actor, on classes a corresponding object etc.) If that's not possible: Exists there some sort of workaround?

4

General Board / problem with aggregations and attributes

« on: December 03, 2003, 04:15:38 pm »

Hi,

I'm playing around with the trial-version of EA for the first time so perhaps I simply haven't understood the "principle" behind some functions. But something with the code-syncronization-mechanism seems really odd from my point of view. Here's a receipt to reproduce the effect:

* create a new project (based on EABase)
* add a new class ClassOne in the Data Model class diagram
* add a new class ClassTwo in the Data Model class diagram
* add an attribute objTwo: ClassTwo to ClassOne
* add a strong aggregation between ClassOne and ClassTwo and name the TargetRole objTwo
* generate C++-sourcecode for both classes
* rename the attribute to objTwoRenamed in the sourcecode
* synchronize the sourcecode with the diagrams
* now only the attributename becomes changed not the named aggregation. So if you generate sourcecode a second time now, you'll have two attributes objTwo and ObjTwoRenamed in your code.

Couldn't you treat named aggregations internally just like another form of attributes (what they essentially are)? Perhaps with a switch in the "Model Attributes"-Section and while importing source, whitch representation should be choosen (aggregation,attribute or both).

Another annoying detail: If I delete an attribute or method in the model and synchronize with the source, the deleted elements stay in the code in the default settings. There are flags for reverse engineering in the Local Options-Dialog. Is there a similar option for forward engineering?

Greetings,
Ferdinand