Re: [oiic-formation-discuss] My perspective. display perferct?

[jose lorenzo <hozelda@yahoo.com>]

--- On Tue, 7/15/08, Bart Hanssens <bart.hanssens@skynet.be> wrote:

> From: Bart Hanssens <bart.hanssens@skynet.be>
> Subject: Re: [oiic-formation-discuss] My perspective. display perferct?
> To: oiic-formation-discuss@lists.oasis-open.org
> Date: Tuesday, July 15, 2008, 4:13 PM
> jose lorenzo wrote:
> > Should ODF (or something/someone else?) state as a
> conformance clause
> > that vendors claiming conformance must allow the
> results of
> > "official"(?) testing to be published. 
> 
> Mm, dunno, I would be quite happy having an
> "official" test suite and
> firm wording in the spec, without going as far as
> trademarks or forcing
> the implementor to publish the results. Is there such a
> requirement for,
> say, PDF/A or Docbook ?
> 
> After all, if the standard and test suite are freely
> available, any
> customer (and large prospects) can test whatever
> implementation a vendor
> wants to sell...

Yeah, I think the leaning by some has been towards there being no official testsuite implementation, but what you suggest is something to keep in consideration. And yes everyone would have access to it too, so big vendors (if not the little guy) would be able to test within their labs.

Another possibility would be for OASIS to certify testsuites that conform to whatever testing prescription get put out. Different testsuites would be useful to different target audiences, for example, in terms of the time it takes to run (overall comprehensive) or in terms of focusing on specific sections of the spec (less time but maybe hit hard some areas germane to specialized limited tools).

Also supportive of not having a single official implementation and/or for certifying multiple ones is that there might be competition among test suite implementations, even within a particular class of suites, with the "best" suite title changing hands over time. By certifying multiple ones, the user has a choice, where some particular choice might be best for their needs, but any choice would be known to essentially be correct.

There may also come to be profiles (each with own conformance criteria). Even if not, the OIIC TC may find it wants to recognize categories of testing. I think this would imply that it would not make sense to have a single official test suite. In particular, no single group may care (find the resources, etc) to build a full test suite for every component.

And in lieu of OASIS doing the certifying/testing of the test suites, they might certify those that do (so OASIS would add a "management" layer).

[Some but I don't think all points above have been brought up before on the list.]