xml-dev - Why Bother With Standards? (Was RE: RELAX to ISO)

["Bullard, Claude L (Len)" <clbullar@ingr.com>]

Right.  

The lunatic left wing fringe is a reference 
to how the earliest advocates of using SGML for 
hypertext were referred to by some in the print 
community.  We were told over and over again that 
it was "impossible" to create hypertext systems 
with SGML.  That was still being asserted as late 
as 1993 in influential circles and meetings.  Goldfarb 
and Rubinsky took a program called IADS to a 
NIST conference to prove otherwise.  It silenced 
some critics but others asserted it proved the first 
assertion because it needed a thing called a 
"stylesheet" to interoperate and reserved 
four element types for its own use.  How dastardly... 
If it has used processing instructions (did in the 
original design), they would have brought out torches, 
tar and feathers, and ... well, you've seen mobs.

Then came HTML with many tags, no stylesheet, no 
content independence, just a presentation with 
hyperlinks, and what did they say?

They said "It is a standard!"  But it wasn't.  It 
was just on a lot of platforms, was free, had open 
source implementations, and so on.  It moved 
like settlers across the midwest, colonizing, 
wresting natives from the lang, making claims, 
building sites, and finally, becoming ... defacto.

There is rhetoric and labeling, everso politically effective, 
then there is freeware distributed as fast as word 
of mouth can couple with the rhetoric and labeling,
then there are the hard signatory, authoritative 
agreements that make up the domain of "real standards". 
In that domain, terms like "normative", "informative", 
"annex", "standard", "conformant", "compliant", and 
"specification" have precise meanings free of rhetoric 
and implementation bias. 

In that world, XML is a subset of SGML.  ISO can pick 
up XML and manage the standard document and normatively 
reference the W3C specification, but in effect, they 
already do that.  It is ISO 8879 plus all of the ammendments, 
annexes, TCs etc.  If someone is poking around in there, 
they have a very hard problem to solve.  They care a 
great deal about something.  

Then there is "running code".  Well, it colonizes 
but it does not standardize.  It occupies space 
on the harddrive, but it is only as public an 
agreement as the platform owner is willing to 
demonstrate it.   

Who cares?  Why bother with standards?

Fact is, anybody can do anything anytime with any of 
the published document contents if they relabel what 
they do.  That is the game of market, switch and bait, 
embrace and extend, the dot In Com, and so forth.  Unless 
it is supported by running code on some very large number 
of platforms, it is irrelevant to the user.  Unless it can be tested 
and proven to comply and conform to the specification or 
standard, it is irrelevant to the organizations that claim 
the right to create such for some polity.

But if there is no credibility in the claims, no honor 
to agreements made in public or private, then even the 
irrelevancies are irrelevant because the recourse as 
in world events today is to return to tossing stones, 
returning fire, and erecting barriers to cooperation, 
because competition for space on the platform is the 
one real fact that cannot be bargained away.  The winners 
of that process will be the companies and individuals 
who figure out what customers want and need and provide that. 

Each polity will commit to leaders and if lead badly, 
will not be found except in the niches.  You will 
find it comes down to individuals who stay the course 
and survive.  Gritty work.

Len
http://www.mp3.com/LenBullard

Ekam sat.h, Vipraah bahudhaa vadanti.
Daamyata. Datta. Dayadhvam.h


From: Rick JELLIFFE [mailto:ricko@geotempo.com]

"Simon St.Laurent" wrote:
 
> Maybe someday I'll run into an XMLer who cares deeply about SGML who
hadn't
> used SGML before the advent of XML, but I've yet to find that creature.
If
> that makes me lunatic left wing fringe, that's fine too.
 
I think Len is refering to those for whom "XML is SGML" is an
inconvenient barrier to pet changes rather than the great unwashed masses
for whom it
is merely a remote and uninteresting factoid.   If they care about XML they
care about standard generalized markup languages, and if the capitilized
source doesn't affect them but gives refuge to civil servants, who cares?