Re: [regrep-semantic] FW: METEOR-S Web Service Annotation Framework Release

[ewallace@cme.nist.gov]

Regarding the recently announce METEOR-S tools, David Webber wrote:

>These guys are truely geniuses - they are claiming compliance with 
>BPEL 1.1 when not even the BPEL team members know what compliance or 
>conformance to BPEL 1.1 can be defined as or even tested and assurance 
>or verification given by whatever means
>(see postings on BPEL list last week).

It is quite common for organizations to make such claims early on
just because the validity of the claims are impossible to verify.
However, in ISO parlance "compliance" means nothing more than a vendor 
claim, whereas "conformance" claims require normative metrics for 
testing.  So I guess this makes the authors of these claims doubly
clever, although I have yet to find any claims like this in any of 
the METEOR-S material that I skimmed this morning.

>But apart from that the notion is interesting - but clearly in the 
>research domain - and may provide some useful ideas for salvaging into 
>our ebXML Registry work -but one senses that all this is merely a very 
>complicated way of doing what registry can already do easily with 
>classifications and associations...and reference to BPSS.  Some OWL 
>does add value beyond that however - but again one needs to understand
>the business goals before applying technology indiscriminately.

This work is clearly research grade at the moment (and for the 
foreseeable future IMHO).  Concept matching and dynamic service 
composition are in fact popular areas of research at the moment.
But I don't see how either can be handled *at all* with the current
regrep.  An ebRIM conformant repository could be a component in a 
system that would do these things, but it would just serve up inputs
and record vetted results, not provide any inferencing.  An SCM enhanced
repository could change that in the future if it supported some of the 
classification based query scenarios that we have discussed.

-Evan