[Uml-user] Bugs and further development of umbrello

[dr.andreas.schenk at gmx.net]

Hi,

I started to look into umbrello last weekend. First of all, I should say that 
I was impressed. I didn't expect to find so much functionality in this 
project. Thank you for providing this application as open source!

Unfortunately, much work remains to be done, i.e. there are also some bugs to 
report. I am using release 1.3.0 for SUSE 9.0 with SUSE 9.0. This release 
cannot be used to do real work. It crashes too often. Hence, I tried to build 
1.3.2, but it requires Qt 3.3, which is not part of SUSE 9.0. Is it possible, 
to get this release for SUSE 9.0? I don't want to go back to an older 
release, since it is missing to much functionality.

Next I have to report some bugs in detail. I am using java as the only active 
language. 

I created a class of the form a::b::c::d::e::Bank with a composition 
association to the class a::b::c::d::e::SecuritiesAccount. In the generated 
source for the class Bank I find the statement

import a::b::c::d::e.SecuritiesAccount;

(note, that only after e there is a dot, as required by the java syntax) 
instead of 

import a.b.c.d.e.SecuritiesAccount;

Then I would like to define a static operation in the class Bank, that returns 
a list of all Bank objects. The signature should be something like (in java)

public static List getBankList() {...}

or, in UML an underlined (since static class method)

+ getBankList() : List
--------------------

where List is the java class java.util.List. If I do so, the code for the 
class is generated wrong: umbrello generates

public List getBankList()

without the static keyword.  On the other hand, I miss the required 

import java.util.List;

I don't know, how I can specify a java type of some library, so that umbrello 
inserts the correct import <<type>>; statement.

Maybe some of these bugs are already fixed in a newer release?

In the long run I would like to model J2EE applications with UML. I don't 
know, whether the functionality of umbrello is sufficient for these kind of 
models yet. In case I find some points for improvement, are you interested 
that I share my experience?

Regards,

Andreas Schenk

P.S.: I am also interested in UML 2. There was already an E-Mail concerning 
this topic on the list today.