The submission itself looks similar to what I looked at before. Extremely high level; not much there for me to react to yet - they are still wrestling with (from my practitioner's perspective) non value adding meta-meta work. (Yes, I suppose that being able to validate that if an XML type is acyclic then its derivations must also be is important... has to do with the graph-centric nature of internal IT data for those of you who are curious - but man, this is deep foundational work.)
Somebody please wake me when we start seeing some attempts to actually define service semantics.
I'm just sorry they didn't submit this to the OMG instead, but with Microsoft on board no way that was going to happen, and I'm also not sure how interested the OMG would be in this.
All I can say is that I want to model these services once, in a common environment with my software system construction. Is that so much to ask? Now we will need a translation from UML deployment diagrams into SML - no big deal, I suppose - preferably through some DSM workbench that supports both UML and SML. The SML models in turn will feed my release management system and CMDB.
Would really like some discussion on this one.