on UML, target audience, document structure ...

[Ken Laskey <klaskey@mitre.org>]

I'm currently trying to catch up on email and realize I haven't gotten  
SOA-RM email for a couple days, so please excuse if this is covering  
old territory.  (Also, if this gets a bit long in the tooth, please  
just skip to the bottom and see another reference I suggest we look  
at.)

There have been many thoughtful posts on these subjects.  I agree with  
Rex Brooks that we need to go about this systematically because in my  
experience if we don't, we will redo much and revisit issues which  
should be long since buried.  Rebekah Metz brought up the numerous  
potential audiences and while we can't be everything to everyone, if we  
can't clearly describe our results to anyone but the experts, then we  
miss our opportunity to have a real impact.  This said, I agree with  
Greg Kohring that we need clear unambiguous (possibly multiple)  
presentations, and the presentation will need to be clear and  
appropriate for each audience we choose to address.

So I suggest the following:
1. Let's begin by figuring out what we have to say.  This should start  
with collecting (and challenging) our assumptions about what an SOA is,  
what it does (or helps us do), and why we care.  Don't worry at first  
how to capture it formally - let the initial capture be the challenge  
for our editing crew.

2. As we begin to understand the domain, let's consider how it  
naturally partitions itself.  Part of that will be technical, part will  
be the audiences for which we'll need to emphasize (or expand upon)  
which aspects.

3. Then let's figure out the best way to present and refine our ideas.

For those that cringe that we'll wander forever, that will only happen  
if we always express new ideas without giving any context for what has  
already been said.  So when you bring something into the conversation,  
say what it connects to, what it adds to, or what it is juxtaposed  
against as a different way to view the problem.

Do I guarantee this will converge?  Well, let's say I'm cautiously  
optimistic.  What I do know is that a false sense of structure does not  
make hard points any easier and can deter making real progress.

Enough on general philosophy, as a concrete contribution, I suggest  
looking at the OWL-S work  
(http://www.w3.org/Submission/2004/SUBM-OWL-S-20041122/).  I am not  
endorsing their solution but they have thought a lot about the problems  
and needs for describing services and making services work.  Their  
figures use a concept map (directed line graph) approach, so also look  
at this as an example of presentation and see how you like (or dislike)  
it.  I haven't yet read the companion document is OWL-S' Relationship  
to Selected Other Technologies  
(http://www.w3.org/Submission/2004/SUBM-OWL-S-related-20041122/).

Enough already :-)

Ken

------------------------------------------------------------------------ 
------------------
Ken Laskey
MITRE Corporation, M/S H305     phone:  703-883-7934
7515 Colshire Drive                        fax:        703-883-1379
McLean VA 22102-7508