[PEAK] XMI 1.2

[Phillip J. Eby]

At 04:19 PM 3/23/04 +0100, Sofia Papagiannaki wrote:
>Hello,
>I'm interested in parsing xmi files produced by Poseidon.
>I tried to use peak but I get some exceptions.

Unless I see the exceptions, I have no way to tell what's happening.


>  I guess
>this happens because Poseidon saves diagrams in XMI 1.2
>format?

Does it?  What version of UML does it use?  Can you give me a URL to 
download the file you're having problems with?



>   I'm specifically interested in the peak XMI parser
>because it is written in python, so I would like to know when
>do you plan to support XMI 1.2.

I don't have any immediate plans to work on this.  But if you look at the 
TODO in peak.storage.xmi, you'll see this:

         XMI 1.2

             XMI 1.2 is mostly a simplification and clarification of XMI 1.1:

                 - Encoding of multi-valued attributes; note that it is not
                   permissible to have a value for a feature both in an
                   object tag's attribute and in the object's contained tags.

                 - "Nested packages may result in name collision; a namespace
                   prefix is required in this case."  Need to review EBNF,
                   and "Namespace Qualified XML Element Names".  This may
                   require metadata support on the writing side.

I think it's unlikely that either of these differences would be causing 
exceptions.  Indeed, my first guess is that the issue is more likely to be 
with what version of *UML* is being used, and not what version of XMI.  XMI 
is just an interchange format -- a standard way of mapping a model onto an 
XML schema.  The model, however, can be UML 1.3, UML 1.4, or UML 1.5, for 
example.

Indeed, the Poseidon site makes mention of UML diagram exchange 2.0, which 
I'm not at all familiar with, so I don't know if maybe that's causing a 
problem.

So, I really can't help you much unless I have a file to look at; 
preferably something very small and short, but that still causes you a 
problem.  I'd prefer you not send it to the list, either, but instead give 
me a URL where I can retrieve it.  Thanks.