Re: Namespaces, the universe, and everything

Dave Hollander wrote:
> 
> This is so true. The web's success provides it the high ground.
> I must argue that the SGML lessons will help the web, not the web
> must help SGML.  Since we are all moving toward a new technology,
> any point will lose if it is pressed based solely on the premise
> that the SGML community knows better.  If that argument is used enough,
> the SGML foundation of XML may become suspect then challenged.

That was precisely the argument Henry Ford gave his managers for 
sticking to the Model-T.  It was wildly successful, everyone had one, 
cheap, easy to fix, and who needed anything other than black.

Call it rhetoric if you like.  It is history.  GM took over 
the leadership in the market with a component-based design 
to stratified markets of buyers.   Ford Motor Company never 
recovered its dominance.

Henry took to shooting at his workers.

The SGML community knows SGML.  That made HTML trivial to 
learn and use.  It is also the reason we are here trying 
to improve the potential for markup applications on the 
Internet.  We defend tools and approaches which we understand, 
use and use well and have for a number of years before the 
Web.  We understand the Web as well.  It is HTTP: 
the poor man's protocol which for the most part, is a 
DEC 9600 tape format a la 1840 put in a wire with two methods 
and MIME: types by file extensions.  Wow.

If the only argument being presented was that 
"we know better", you might have a point, but it isn't. 

The SGML foundation of XML is why many of us agreed to work for free 
at night and in our spare hours on this worthy project.  To 
eliminate that reason is to also betray the trust of this 
working group to those whose names do not even appear on 
the membership list.  That is without honor.

This discussion will lead to schism.  One group of consultants 
will build something XML-like for one of the two 
browser vendors, and the other will try to fix SGML.
XML without the legitimacy of SGML will fail.  SGML 
without a subset for the Internet will yield to PDF.

I think we can do better.

Len Bullard

Received on Thursday, 19 June 1997 23:06:03 UTC