[
Lists Home |
Date Index |
Thread Index
]
On Mon, May 19, 2003 at 10:42:25AM -0400, Jonathan Robie wrote:
> >Poisoning the well is [...] [a] sort of pre-emptive strike.
Rhetorical gaffes aside, there is a point lurking underneath the surface:
de facto vs. de jure standards. A goodly portion of the internet was
built on de facto standards, *not* de jure standards dictated by standards
bodies. (Remember the ISO OSI seven layer taco? X.400?)
> XML, XPath, XSLT, DOM and CSS were all developed by standards bodies, and
> have been widely accepted. RELAX NG has not achieved widespread adoption,
> but the standardization of RELAX NG core has not hurt its design.
The issue of de facto/de jure standards is close to becoming a permathread
in one form or another. A key reason for this is that the terms of
engagement are quite murky; is XML a de jure standard, or is it a de facto
standard that was simply codified by the W3C? Ditto XPath, XSLT, any
other "standard we like" (FSDO "we").
One point that never seems to get any attention is that de facto standards
do not automatically equate to "goodness", just like de jure standards do
not automatically equate to "clunkiness" or "badness".
If XML is to be considered a de facto standard (bear with me here...) then
it must be evaluated against comparable "de facto standards" in its realm:
TeX, RTF, YAML, Pod, Stuctured Text, Restructured Text, Property Lists,
.ini files, CSV files, etc. In that company, it's rather clear that
there is nothing intrinisic to de facto or de jure standards that make
them either good or bad.
The one thing that *is* clear from Jim Waldo's piece is that premature
standardization, like premature optimization, is an unnecessary evil in
this business. Sadly, that point is hidden in an attempt to disprove
the "all standards are good, and all good things are standardized" myth.
Z.
|
