[Uml-devel] usability suggestions

[Andrew Sutton]

> but for a 'ground-up reengineering efford' you need a critical mass that is
> hard to assemble. Wouldn't it be possible to figure out a continues path
> from the current design to one where your omf could be dropped in ?

it's hard to say if there's a direct (or even roundabout) path from the 
current state to a place where the omf could be dropped in as the underlying 
data model. it'd be great if there was, but i'm not so sure.

even if there were, there has been some discussion, some of it going back over 
a year, about some other issues with the umbrello organization. one of the 
other wishlist items was better modularization - specifically using kparts 
for each of the various diagram types. another was a better rendering canvas. 
these items alone would require a fairly significant rewrite and if they were 
to happen, that might be a more appropriate time to consider using the omf.

you might also consider that umbrello may choose to adopt the next version of 
uml (2.0) in the future. of course, i had planned on building that for the 
omf at some point too.

of course, the sooner the better :) i'd love to get some real users!

andrew sutton
asutton at cs.kent.edu