[
Lists Home |
Date Index |
Thread Index
]
Anyone can make their own club with a charter and
founding members. It's a free country. So no, I have
no more objections to this than to the skunk works efforts
where the original authors of XML 1.0 go off and create
their own profiles. Same game; different scale.
I note that the WSIO papers thus far don't mention XML Schema as one of the vital
web service standards although I thought they would. Since co-occurrence contraints
either are in the schema or have to be in the business logic, that would also affect
the web service capabilities with regards to which specs are in the profiles. Profiles
sound like a good idea until you acutally write them and discover how QOS is affected
by the profile itself. They have some serious work and they are just as likely
to get caught in the mire unless they already have an answer and this org is one
designed to promote a fait accompli. Again, no objections on technical grounds; just
yetAnotherBunchaStuffToRead.
Did anyone get a chance to look at Tim Bray' et al's Skunk Work paper on XML over the
weekend? It seems to me that of all the current threads, this one has the most
impact on every other topic. That is, if we really want a layered architecture,
some options in the schema space, etc., a slimmed down XML 1.0 profile that
takes out DTDs so they can be moved to another layer is the right place to
start. I'm not as happy that namespaces are there but am willing to entertain
arguments otherwise. Can namespaces and alternatives such as John
Cowan's reworking of Archtectural Forms work together? It seems they can
but for what functionalities?
In other words, will a slimmed down XML 1.0 compliant profile eliminate
options such as Archtectural Forms by insisting that namespaces are more
core than DTDs? I can see that from the point of view of disambiguation
but not from semantic associations.
len
-----Original Message-----
From: Mike Champion [mailto:mc@xegesis.org]
2/11/2002 9:15:41 AM, "Bullard, Claude L (Len)" <clbullar@ingr.com> wrote:
>This is a push to get to production with "good enough" specs
>and tech.
Does this seem like a problem to you? Whatever one thinks about the
current Web Services hype vis a vis what we think we've learned about the
web architecture, and whatever one thinks about "good is best" vs "best is
best", what options do the WS-I participants have other than working
together in something like the WS-I? All they're doing (according to the
interview) is define profiles of the various specs (at least Lewin doesn't
call them "standards") that people want to actually use, build best
practices guidelines for using them in an interoperable way, and build
test suites.
The alternative of waiting for the "real" standards bodies (or
theorists/researchers) to do this is simply not an option that one could
advocate and keep one's job, I would guess. Hype is a cruel master, and
it is applying the whip and saying "row faster." Most of us here know that
the SOAP bubble is inflating faster than the reality behind it. If the Web
Services advocates pursue the vision (theoretically flawed though it may
be) in a coordinated way, at least we will KNOW what when wrong when the
bubble bursts. If they do it willy-nilly in a quasi-proprietary way, we
won't know (in a few years) if the bubble burst because the idea was
doomed or because the execution was shoddy.
Somewhat more optimistically, the Web Services vision has a better chance
of becoming a near-term reality with coordination being done by an
organization with a totally pragmatic charter rather than one of defining
long-term standards or "leading the web to its full potential." It's quite
possible that necessary adjustments to the vision (such as deprecating RPC
and using more scalable and robust message exchange patterns) can be done
without throwing the baby out with the bathwater, but this requires a
process with much tighter feedback loops than the W3C or ISO has these
days.
The Web needs a) short-run pragmatic coordination among fierce rivals; b)
longer-term architectural planning and integration; and c) Rigorous
standardization of the technologies that prove their value. Perhaps in
the SGML era, the ISO did all three. The W3C evolved to fill role a) in
its early years, then matured to focus on b). That leaves role a)
unfilled, and the WS-I was invented to fill it. Of course, people who
want to avoid the chaos of living in Internet Time should stay away from
the WS-I until it produces something, and support "real" standards efforts
from in a revitalized ISO activity and/or maturing W3C activities.
|
