[
Lists Home |
Date Index |
Thread Index
]
4/24/2002 6:07:49 PM, "Evan Lenz" <evan@evanlenz.net> wrote:
>Are you referring to XML 1.0? I always considered it to be a good de jure
>codification of de facto best practices in (i.e. most common subset of)
>SGML. In that sense, Michael's characterization seems better than "real new
>science", let alone a joke...
That's exactly the way I view the history, hence my post.
XSLT is another interesting case ... one could hypothesize that its success stems in part
from the fact that it is "a DSSSL subset with angle brackets" (at least that's how it was
explained to me in about 1998). The DOM Level 1, likewise, was more than a "codification of
de facto best practices", but built heavily on existing practice in both Dynamic HTML
browsers and the scripting languages built into various SGML products.
I am quite willing to concede the point if I'm clearly wrong, but I believe that the history
of the W3C specs, the hits and the misses, strongly supports Gosling's 1990 Window of
Standardization hypothesis, tongue in cheek though its presentation may have been.
http://java.sun.com/people/jag/StandardsPhases/index.html
In a nutshell, it predicts that successful standards require a situation where:
1 - The technology in question has become understood, thus less "interesting" to specialists;
2 - The political interest in the technology is high enough for it to be worth investing in,
but not too hot as to prevent constructive negotiation.
Here's my examples/disconfirmable predictions: XML 1.0, DOM Level 1, SOAP 1.1, and probably
XSLT were right in the window; Namespaces and W3C Schema are less "successful" (if the amount
of whining on xml-dev is an indication!) because they were too technologically "interesting";
the the various post-WSDL 1.1 web services specs are not likely to become "standards" until
the political interest cools down quite a bit; the SemWeb will not be a success until the
technologies behind it become a LOT less interesting than they have been for 20 years or so;
and XQuery will be a successful standard to the extend that they do NOT try to do "real new
science" by committee but rather stick with some less "interesting" technologies.
|
