Re: Umbrello and XMI

Jonathan Riddell <[email protected]>
Newsgroups gmane.linux.umbrello.user
Message-ID <[email protected]>
On Fri, Apr 09, 2004 at 04:02:14PM +0000, Anna Persson  wrote:

> 1)Is there any plans to support newer versions of XMI (like 2.0) in Umbrello?

We would like to support XMI as the standard for UML as XML but it is
a very complex standard (and XMI 2.0 will probably be even more
complex) and I havn't found any guide which makes it easy to follow.
So we don't follow the standard very well and are unlikely to support
2.0 once it comes out.

> 2)Do you think it is critical that Umbrello can be integrated (can exchange information) with other modelling 
> tools?

It would be very nice if it did but at the moment that just isn't very
possible, we don't have the people power to make it happen.

> 3)Some modelling tools doesn't support ?general? import/export of XMI files. Instead they support import of XMI 
> files from a number of specific tools and they export XMI files for specific tools (eg. Rose XMI) . Is there any 
> plans to suppport tool specific XMI import/export in Umbrello as a complement to the current ?general? 
> approach?

Same as above.  And I wouldn't look too favourably on supposrting
proprietry extentions to standards.

> 4)XMI is an ?open? standard that can be ?adjusted? by tool developers implementing it. This means that 
> interchange of XMI information between two modelling tools may not work, despite that the two tools supports 
> the same version of XMI. Do you think that XMI as on open standard currently is succesful? Or does the XMI 
> standard have to be ?less open? in order to actually work?

It doesn't seem to be very sucessful in that all UML tools supposrt it
differently.  

> 5)Many modelling tools can just handle one specific version of XMI, which limits the number of XMI files that can 
> be imported in the tools. Why is it like this? Is it hard to support several versions of XMI in the same tool? Or is 
> the tool developers more interested in the export function rather than the import function?

It's hard to support one version of XMI never mind follow other implementations.

> 6)My study has shown that exchange of XMI information between modelling tools are really troublesome and I 
> know that others have experienced the same problems. Are these problems related to insufficiences in the XMI 
> standard or to insufficiences in the tools? Or both?

Possibly to insufficiences in XMI and certainly to the complexity of
the standard.

> 7)What do you think about the future of XMI? 

XMI 2.0 will mix SVG with XMI I believe?  SVG is horrendously complex,
I havn't seen a complete implementation of it yet, same with XMI.
Putting them together sounds like a nice idea but it won't be simple
to implement.

> 8)Are there any alternatives to XMI (apart from solutions based on repository technologies)?

http://www.yy.ics.keio.ac.jp/~suzuki/project/uxf/

I havn't look at it at all.

> 9)Is there an  interest in information exchange between tools from users of modelling tools?

Yes, only yesterday a developer was saying he'd done some UML diagrams
and then been told he was using the wrong tool.  A standard file
format would be great.

Jonathan Riddell


-------------------------------------------------------
This SF.Net email is sponsored by: IBM Linux Tutorials
Free Linux tutorial presented by Daniel Robbins, President and CEO of
GenToo technologies. Learn everything from fundamentals to system
administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
lmpx.com only provides a reader for public news (NNTP) servers. It is not affiliated with the servers or forums shown here and is not responsible for the content of articles, which is written by their respective authors.