Lists Home |
Date Index |
Don Demsak wrote:
> I don't think it would be beyond some of the business on the committee
> to purposely slow down the recommendation process in order to give
> their companies time to catch up to level of support Microsoft has for
> XQuery. So which one do you think will be released first Longhorn or
> the W3C's XQuery spec?
As one of two staff contacts on the XQuery WG I can say I've seen no
evidence of Microsoft attempting to slow down the process -- if anything
they've pushed hard to go faster.
As far as I can tell, some of the other implementations -- both open
source and proprietary -- are at least as far advanced as anything
Microsoft has. Although Microsoft has demonstrated some XQuery support
in SQL Server, IBM and Oracle have also made announcements, and if you
take the time to look at the public XQuery page  you'll see links to
a number of implementations of the current draft, including several that
have open source licences. Galax, Saxon and Qizx/open are three that
I've tried recently and seem reasonably complete, but there are
> One of my major complaints with the W3C (and other standards
> organizations) is that they just produce standards, not
> implementations of the standards.
Are you willing to help finance the extra work?
I understand that software
> companies have a vested interest in releasing products
> according to a specification, but without having a publicly
> accessible implementation of the spec to work with during the
> draft process it makes it very difficult to create test
There are several open source implementations. Two of the ones I
mentioned above are worked on primarily by people who are participants
in teh XQuery Working Group.
> My idea is to marry a standards organization with an open source
> community (think of merging the W3C and SourceForge), but put a hard
> division between the two groups. In order to prevent intellectual
> property leakage from the businesses on the standards org side to
> the open source side, individuals from one side can not work on the
> other (for a given spec). This way there is a living example for
> the standards group to work the bugs out of (before it becomes a
> recommendation), and should streamline the specification process.
There are implementations in multiple languages on SourceForge already.
> To make this work, the open source project leaders would need
> to have very good access to the working draft committee.
Such as being participants, for example...
We actually want people to be able to implement our specs from
the specifications, rather than by telephoning the authors of the
specifications... as a result, it's very encouraging to see a number of
implementations (both open source and closed) from "outsiders", and to
see already a good degree of interoperability, although I should stress
we've not yet defined what it means to be conformant, nor been able to
release a test suite beyond the sample use cases.
> In a perfect world, there would be at least 2 open source
> (.Net and Java), and even the code based on
> rejected implementations would be available for all (which is
> something some of the MS MVPs in XML have been asking of MS
> for libraries that were abandoned).
We don't reject implementations. Something that doesn't conform might
still be useful, although erhaps less popular.
The W3C doesn't do conformance testing of implementations. That's a
pretty expensive thing to do, so we'd probably have to charge for it
(say, $100,000 per product?) which would obviously exclude many open
Don, please do take the time to read the W3C's public Web page for XML
Query  before writing things that are simply not justifiable, and are
easily refuted with even a very cursory inspection.
W3C XML Activity Lead, http://www.w3.org/People/Quin/
Also, alternate staff contact on the XML Query Working Group