Welcome everyone ... including me!

I had asked for this forum, but only chanced to notice it at the bottom of the list this morning.

I'm not sure why it is here, under PSDN, instead of under OpenEdge ... seems like it would fit better there.

But, whereever it is, let's talk.  The concept is to look at the potential and real work being done with modeling for ABL.  To me, this divides into three broad categories that have their obvious interrelationships, but which tend to correspond to three unique orientations.

The first is the use of modeling as a way to build new applications.  This is the classic OOA/D which has become fairly prominent in 3GL OO development.  Use of modeling varies considerably from those who use it mostly as a white board to get concepts clear before coding, through those that use it to develop detailed specifications, but still do the coding manually, to those who use MDA (Model-Driven Architecture) to actually generate some or all of the new system.  A flavor of the latter includes the idea of Executable UML in which one actually tests the software based on the model without generating any end system code.

The second is the use of modeling for analysis, i.e., to help understand and maintain an existing legacy application.  Here the trigger is often having a legacy application that has received years of modification with no documentation and which probably had minimal documentation to start with.

And the third is transformation, i.e., taking a legacy application and transforming it into a more modern architecture.  This can involve some amount of analysis as in the second category and it can involve OOAD and even MDA as in the first category or it may only involve simpler versions of these efforts.

iMo is perhaps the best known of the efforts in the first category ... best known in that we know they are doing it, but not best known in knowing what they are actually doing.  It does seem to apply only to new software development and does seem to use MDA, but those of us on the outside have little visibility to the actual models and techniques.

My own ABL2UML effort (http://www.oehive.org/ABL2UML) is an example of the second category ... there may others and I am pretty sure PSC Professional Services has done something in this area, but I think ABL2UML is the only one which is publicly available.

PSC Professional Services is the most conspicuous actor in the third category, e.g., Phil Magnay's recent talk at Exchange.  But again, details are limited thus far.

So, let's hear from those who are interested in or actually doing something with modeling.  Tell us what you would like to do, what obstacles you have encountered, and what successes or tips you might have.  There are several fairly obvious topics we could talk about.  Among other things, I would like to see if we couldn't get some movement going on adopting some standards so that tools and techniques developed by one person might be useful to another.  I will certainly be monitoring the forum and I think Phil will be as well ... to the extent he can with his busy schedule.