[w3c sml] Bug 4675 - SML-IF conformance concerns

As requested in our last meeting, I am re-sending my email that expresses my concerns about our conformance criteria. I also dug up Sandy's latest proposal (note that Bug 4675, comment #24 is NOT correct - the attachment does not reflect this latest proposal). I created a new document that includes only the parts that affect Bug 4675 - so the latest proposal for 4675 is attached to this email.

I wish to restate that I don't think the WG has spent enough time discussing this latest proposal in the context of our goals for interoperability/conformance to take a vote on this now.



--
ginny

-----Original Message-----
From: public-sml-request@w3.org [mailto:public-sml-request@w3.org] On Behalf Of Smith, Virginia (HP Software)
Sent: Tuesday, December 11, 2007 3:39 PM
To: public-sml@w3.org
Subject: [w3c sml] SML-IF conformance concerns


In addition to my already expressed concern that having 2 levels of conformance dilutes the interoperability that SML-IF is trying to provide, I have concerns about the SML-IF conformance criteria proposal.

The latest proposal is that full conformance requires every definition and instance document to be embedded in the IF document:
1. Embedding all definition and instance documents can create very large IF documents.
2. It seems unnecessarily inflexible. I can see scenarios where an SML model uses the URI scheme and some documents, particularly schema documents, will be available at a 'well-known' location (i.e., URI). It might not be necessary and or even desirable to embed the schema document in the IF document. But, in this scenario, the IF document would be considered only minimally conformant which would definitely not be an indication of its interoperability.

In addition, I would like to see the WG spend time working through the scenarios proposed by Sandy. For example, are signatures on individual documents likely to be an issue or just an interesting use case? Or, perhaps, a component providing a document to be part of an SML model would be expected not to sign references. And the question of "Different vendors could agree on a new scheme and start using it before it's adopted by the next version of the spec. Do we want to clearly disallow that kind of behavior?"  hasn't been answered particularly in the context of the conformance criteria. In addition, the connection between interoperability and conformance has been mentioned but no substantive discussion has taken place to explore this.

I know we are coming to the end of the year and there is pressure to wrap up but I must say that I am uneasy with the conformance criteria for IF and my feeling is that the WG needs to spend more time discussing conformance before making a final decision.


--
ginny

---------------------
Virginia Smith
HP Software / BTO R&D
916-785-9940
8000 Foothills Blvd | Roseville | CA 95747
www.hp.com/software