Sorry Sam,
That wasn't a go at v4.0
I have 80+ analysts here that we have to introduce to UML 2, have them trained and modify the ISD processes to use UML2.
In addition to that, I have to somehow validate XMI for UML2 and check that we can adequately interface system models with projects that use "a certain other" UML modelling system.
My
optimal path would be to have small groups of analysts
1. investigate the benefits of UML2 to their systems i.e. use EA4 as a plaground rather than the standard tool issue.
2. estimate the effort required to convert their modelling to UML2
3. migrate systems on a controlled basis with oversight from the program office.
We have "just" stabilised their use of UML at the 1.5 level - all the process and methods material relates to 1.5 modelling techniques and so does the training material.
So, while I can appreciate the move to UML2 and EA4, I just cant let them "get loose on it" without the proper controls. If we did they would all invent their own techniques, model structures etc etc and goodbye reusability, interchange between groups etc etc all the things we have only just started to measureably achieve.
Bruce
p.s. Under Bruce's 4 stages of technology adoption theory:
[1] "It'll never work here"
[2]"I suppose we could look at it"
[3]"This will save the world and cure baldness as well" and
[4] "Its OK but if it did this this and this it would be better"
For UML 2 I am currently very much at upper level 1 / lower level 2 at the moment. I am waiting on OMGs expecetd release of the spec at end of April to see if they've addressed certain issues.