Lists Home |
Date Index |
To clarify how this whole off-topic evolution thing got started (my fault,
sorry): it was actually a comment of mine on the supposed disloyalty of XML
folks towards the SGML crowd, basically stating that competition in the
technology space is a good thing for the same reasons that capitalism is a
good thing in the technology space. Of course, the result is that 99.9%
people on this list don't believe that capitalism is a good idea (hyperbole
intended), so the analogy didn't have the desired effect, to say the least.
I bowed out of the discussion based on the impression that my ideas,
although they *really* aren't particularly radical, are not widely shared,
at least on this list, and the scope for misinterpretation and
misrepresentation is simply too high. Just to say that this was not
originally related to RELAX NG in any way, shape or form, and I even took
the precaution of changing the subject line from "RELAX NG Marketing" to
make sure that this fact was not missed.
Anyway, I agree with most of your points. XSD has so much in it because of
normal mission creep; everyone has their own ideas and agenda, and I guess
there was a widespread fear that "we better get it into the spec now,
because it's taken three years to get this puppy out, and who knows when the
next revision will be." A perfectly normal, human reaction.
In terms of the future of RELAX NG, I also agree that there are a number of
plausible scenarios. At worst, some of its good ideas will end up folded
into some future version of XSD. For what it's worth, I don't see much of
the agenda of Redmond/Redwood Shores/Armonk in the actual spec. I believe
that if some momentum builds up behind RELAX NG, people will happily adopt
it. This will require creating some of the additional layers that XSD
provides (like inheritance, attribute/element defaulting, key/key ref,
appinfo, etc.) in some clean way. It was also require some basic marketing
efforts such as getting RELAX NG into the standard Xerces-J build, even if
the technical justification for this isn't entirely compelling.
> -----Original Message-----
> From: Mike Champion [mailto:email@example.com]
> Sent: Thursday, March 28, 2002 4:25 PM
> To: xml-dev
> Subject: [xml-dev] RELAX and Evolutionary somethingorother
> I'm a bit distressed to see the list wander off into the Capitalism
> thread, but hey, at least the posts aren't in rhyme!
> Anyway, I think the original issue of the Capitalism thread
> is worth returning to: what do we know about how technology, business,
> society, and/or knowledge evolves that would shed light on whether
> RELAX NG is a viable technology in which to invest time,
> money, and "love"?
> * If you believe some crude notion of international
> capitalism calling the
> shots, it's not worth discussing. The unholy Redmond -
> Redwood Shores - Armonk
> - Davos alliance seems to have already chosen W3C XSD, so get
> to work figuring
> it out!
> * If you believe that "beauty equals truth"
> http://www.nytimes.com/2002/03/26/science/26MATH.html and
> truth will prevail,
> bet on RELAX NG and start writing tools to support it.
> * If you think the world is more complex, and that both the
> private agendas of
> the huge companies (who, let's face it, drove the W3C XSD
> requirements and are
> its stauchest defenders) and the selection process driven by the
> Blind Watchmaker / Invisible Hand will interact to
> determine RNG's fate, we've got a lot to think about.
> First, what are the needs that drove the requirements for XSD
> so far beyond what
> most of us think of as the 80/20 point? AFAIK, it was a) the
> need to map
> complex databases and enterprise application data to XML so
> that data could be
> mapped to exchanged across systems and applications ... and
> b) the need to
> support schema evolution more cleanly, i.e. make minor
> changes to RDBMS or XML
> schemas without breaking everything or forcing a simultaneous
> update of systems
> all over the world. Does XSD really meet those needs? Can
> it be implemented
> consistently to do so? Sooner or later I have to think that
> it will ... either
> the spec will be revised by the W3C or (more likely?) some
> sort of informal
> "schema interoperability organization" will figure out the
> profile of the W3C
> spec that meets the real needs of the Big Guys.
> Second, how long will that take? The hype machine giveth, but
> the hype machine
> taketh away, and one can imagine the "buzz" turning against
> W3C and XSD and
> mega-XML in the meantime. Will that happen? That depends on
> whether you think
> of the creators of buzz/hype as the paid servants of the Big
> Guys or not. I
> personally don't ... I think that hype has a life of its own,
> but since XML/XSD
> is supported by EVERYBODY and nobody but a few lonely voices
> in the wilderness
> is objecting, one can make a plausible argument that this
> time the hype cycle
> won't slide down the trough of disillusionment. We shall see.
> IMHO, in the long run, the Blind Watchmaker beats the
> monopolists and the
> hypemeisters every time. Complex systems (or specification)
> will only survive if
> they are assembled out of reliable components that can stand
> on their own. If a
> spec or a product is propped up only by monopolistic business
> practices or paid
> hypesters, sooner or later the environment will change, and
> systems and
> subsystems will have to survive on their intrinsic merits.
> Herbert Simon's
> essay "The Architecture of Complexity" (I like this summary:
-90.html) is the
classic explanation in the context of computer science.
I don't know whether RNG will "win", or XSD will be cleaned up and rebuilt
of more reliable components, or (most likely) the best features of both will
subsumed by some trancendent synthesis that we can't yet envision. I simply
can't imagine RNG becoming irrelevant, however. It seems worthy of study
support; even if what we use in 5 or 10 years is not called "RELAX", it will
much to the efforts of Murata, Clark, et. al.
The xml-dev list is sponsored by XML.org <http://www.xml.org>, an
initiative of OASIS <http://www.oasis-open.org>
The list archives are at http://lists.xml.org/archives/xml-dev/
To subscribe or unsubscribe from this list use the subscription