You might want to look at OOA Tool. [link] It's free
(for now), and I've used it under Linux for documentation purposes. I
have no affiliation with the tool.

It will handle association classes. It doesn't do qualified
associations, but it has other ways to handle the constraints, that

How to use UML to combine Objects and Aspects.
[link]
Remy Fannader

You can try with MOSKitt ([link]). It's a free tool which
mainly provides:
* Model graphical editors like UML2, Data Base (DB), BPMN, User
Interface, WBS and
* Model Transformations like UML2DB, DB2UML, BPMN2UML, DB2DDL (Oracle,
Postgresql, MySQL)
* Code generation: OpenXava, JPA, PHP.
* Other features: Reverse engineering (Oracle, Postgresql, MySQL),

Similarly, Enterprise Architect also supports all this, but again it's
not free, but great functionality for money IMHO

Phil

MagicDraw does all that, but it's not free.
-Jim
Sent from my iPhone

Responding to Marc...
fUML is essentially the Action Semantics originally brought in with
v1.5. That semantics filled a huge hole in UML by providing a low-level,
standard execution semantic definition. But it is low-level and no one
but a tool vendor or translation methodologist ever has any exposure to

Hi,

I am looking for a UML tool, running on Linux, that is able to :
- handle association classes,
- handle qualified associations,
- run under Linux,
- for documentation purpose (no interaction with Java or C/C++ code
for now).

I already tried ArgoUML, Papyrus and bouml, but non of them seems to

Responding to Michael and H.S. …

As I said with fUML and its operational semantics…

Regards,
Marc-Florian

Von: umlforum@googlegroups.com [mailto:umlforum@googlegroups. com] Im Auftrag
von Michael Jesse Chonoles
Gesendet: Montag, 23. Januar 2012 23:22
An: umlforum@googlegroups.com
Betreff: Re: AW: [UML Forum] UML is semi-formal language Why?

THere is a subset of UML, called fUML that is a formalized subset suitable for precise execution semantics.

______________________________ __
From: Timothy Marc <timothym...@freenet.de>
To: umlforum@googlegroups.com
Sent: Saturday, January 21, 2012 3:15 PM

Hi,
Short answer: Because it has a well-defined syntax (so it is

Responding to Marc...
While I agree for UML as a whole, I would point out that with an MDA
profile on can select a consistent subset of model elements and apply a
rigorous semantics to them. Otherwise it would not be possible to do
full code generation for UML OOA models, and that technology has been

The semantics of executable UML belongs to the programming realm, not the
modeling one. That's OK if models and programs are merged (in other words
if models are just a distraction), otherwise it's confusing (at best).
[link]
[link]

Hi,
Short answer: Because it has a well-defined syntax (so it is
computer-readable), but not an inherent, unambiguous and precise (in terms
of mathematical precision) semantics. The semantics of a UML element is
provided in prose, which is inherently ambiguous.
As a result, a UML model can be read by a computer, but due to the lack of

Hi,
I have just one question about UML.
Why we say that UML is semi-formal language ?
Thank you.

Responding to Lloyd...
I have to disagree with strongly with the first sentence. UML was
developed by the Three Amigos -- Booch, Jacobson, and Rumbaugh -- at
Rational Software. It was a synthesis of their three methodologies for
OOA/D (the Booch Method, OOSE, and OMT, respectively). IOW, it was
designed to be a general notation for expressing OOA/D. From reading

Many would argue that UML must be reorganized along a different perspective
if its use is to be not only more widely spread, but also more effective.
Domain specific languages and profiles will split up the UML, and it's
difficult to see how to simplify it without hampering its qualities.
Actually, the objective should be to get a better match between UML

Responding to Chonoles...
No question that there are multiple OOA/D methodologies and you do need
an MDA profile for each one. But those profiles typically define by
subset exclusion (e.g., the UML subset for translation OOA does not
employ 'protection', dependency associations, etc.). But I don't think

(Please accept our apologies if you receive multiple copies of this
message)

IEEE WETICE 2012
3rd IEEE Track on Collaborative Modeling and Simulation
(Comets 2012)

in cooperation with
AFIS (INCOSE France Chapter)

I think UML grew up to be a generalized object-oriented modeling language,
and not specifically for Object-Oriented Analysis and Design of software (in
its strictest sense). OOAD as a paradigm has strengths and weaknesses,
changing in various contexts. A practitioner of OOAD focuses upon
leveraging those strengths. But there are other viewpoints.

Hi all,

To be honest, I really appreciate the way UML is constructed, though I think
there should be more information taken into account directly like
requirements etc. instead of being proposed by additional profiles.

Anyway, the metamodel is quite complex, agreed, and some fundamental things
are redundantly realized, but that redundancy allows you to use concepts

The distinction external/internal can introduce serious misunderstandings
when applied to use cases. Since UCs are meant to represent an external
perspective, the expression "internal use case" is an oxymoron. The
so-called "internal" UC should appear as an analysis model. It may also
appear as another external UC when set as the requirements to a subsystem.