RE: [oic] Testing Approach Decision #1: What Classes of Products Are To Be Tested?

["Dennis E. Hamilton" <dennis.hamilton@acm.org>]

OK, now that I see where the disconnect over Classes of Product arises, I am
happy with the notion that investigation of interoperability answers these
questions differently than dealing with measurable conformance (with respect
to the specification as it exists).  

I don't expect the floor to move any time soon (i.e., before ODF 2.0) until
an interoperable floor is well-established, and that may actually only be in
terms of a profile.  It may actually be too complicated to bake something
like that into the specification, decontextualized from some agreed domain
of interoperability.  (I do think that the conformance language in the
specification can be much tighter, but I don't think that will move the
floor very far in ways that we might otherwise suppose.)

So, to be clear, we need to decide which flavor of the question we answer
first.  For the one related to conformance, we need but deconstruct section
1.5 of the ODF 1.1 specification.  There are consequences elsewhere, but 1.5
is going to limit how much we can infer about classes of product in the
conformance sense.

After that, I think almost everything we do (considering where the floor is)
beyond understanding what some of the foundation specifications do or do not
required (XML 1.0, XML Namespace, IETF IRI and URI, Zip, etc) and finding
out what implementations do with the various features of conformant ODF
documents.  This Venn diagram is going to have a lot of overlap with
identifying and testing for where there is interoperable support and not.

Um, I am on a break right now, for almost a full week, so my silence will
not signify lack of interest.  I think I will plod along with the Wiki only
over that period.

 - Dennis

-----Original Message-----
From: Hanssens Bart [mailto:Bart.Hanssens@fedict.be] 
Sent: Friday, January 23, 2009 07:31
To: robert_weir@us.ibm.com; oic@lists.oasis-open.org
Subject: RE: [oic] Testing Approach Decision #1: What Classes of Products
Are To Be Tested?

Rob,


> However, ODF 1.1 did not formally define modules.  It would be good to do,
> but would require some refactoring of the schema.  Maybe this could be a
> recommendation of this TC?

Well yes, per charter, we have to advice the ODF TC on profiles etc, so...

Actually, the schema-based-checklist-of-implemented-features we've discussed
during the conference call, combined with the non-normative appendix D
("core feature set"), would be a great start in raising the floor (*)

Probably too late to get it in ODF 1.2, but would be nice to have it in the
next version...


Best regards,

Bart


(*) http://lists.oasis-open.org/archives/office/200901/msg00145.html