Open Standards Processes

[len bullard]

Tim Bray wrote:
> 
> 1. The supposition that the XML process was in any material way less
> open than the SGML process is simply wrong.  

I disagree.  Compare the selection rules for membership in working 
groups.  Who chooses the members of the working group for XML?

> XML was aggressive about
> seeking out invited experts to serve on the SIG mailing list, which
> had very substantial influence on the shape of the spec.  

Yes.

> In particular,
> compare, in the XML process versus any other, the number of people and
> organizations who were actively on top of the spec, really understood the
> issues, and provided serious input.  On that basis, XML's input head count
> is exceeded only by a few of the bigger IETF efforts.

This is true.  The SIG is well staffed.  The best SGML experts 
in the business are there.

Point of history:  When SGML was originally created, there was 
little use of the Internet for list activity of the kind that 
is now possible.  That meant travel and financial support for 
standards efforts that only companies could afford.  So, from 
that perspective, I concede.  As one who encourages lists, I 
do so because I have seen the inherent limitations of airlines 
and hotels as the medium of communication for this work. 

> 2. The supposition that the HTML standardization process can be said,
> in any meaningful sense, to have worked, is simply wrong.  Anybody who
> says this obviously has not tried to implement code that processes
> what the marketplace perceives to be HTML.  

Point of difference:  the HTML process produced a technology, not a 
standard.  But to be more truthful, the Mosaic group implemented a 
technology being argued about by a large list.  Considering the 
average age on that list and the lack of practice, I'm sure it 
was raucous.

BTW:  I was part of the team of Lockheed Martin that did 
implement an SGML and an HTML browser.  I am aware of the 
design's limitations.
  
So, yes, it wasn't perfect technology.  Considering the 
results (The Web), that didn't matter.  When the issue of 
choosing a text design for VRML was discussed, some thought 
that ONLY HTML should be the basis for that.  Some still do.

> This is defined not by any spec,
> but by a basis of functionality that was in Netscape 2, and an unholy mess
> of accretions, with only two companies really allowed to play.  

Not true.  Several companies played.  The W3C source was implemented 
several times.  However, Netscape moved fastest and had the freshest, 
and for that design, most experienced team.  So, they extended 
HTML quickly and cleverly.  Extending an SGML application by 
adding to the DTD is the way its done.  To the lasting chagrin 
of the originators of HTML, they insisted on making a standard 
of it rather than defining it as a tool, which is what it really is.

> I think a
> standard should be something that should serve as the basis for
> implementation.  XML is.  HTML isn't.

HTML is a DTD.  Implementing a DTD IS what you do with it: SGML
application.

XML is syntax unification.  I absolutely agree that this should be a 
standard.  But it isn't.  It's the property of a consortium, to  
paraphrase, "big companies that won't play unless they get their way" 
and that includes insuring a one year lead time on development.  
That is anti-competitive as it gets.  Say what you want about 
the SGML process, Charles Goldfarb is a stickler for insuring 
that this does not happen:  ISO rules backed by a man of 
incomparable commitment to the letter and spirit of the law.

Point conceded:  W3C makes the rules for W3C processes.  The 
chair and all official members must abide by those rules.  
It is the rules I question.  Given ISO rules, the XML processes 
would be different.  

> 3. It *is* the case that the W3C process is, by default, less open
> than some others, in particular IETF.  The hypothesis is that in
> web-space, where there are lots of $N*10^7 bets on the table and
> attack-trained marketing groups behind every bush, there are going to
> have to be some closed doors to get anything useful done.  

That is demeaning FUD.  I doubt there are professionals on this list
that 
cannot be handled by the other professionals.  Offlist is another issue.

> 4.... Such memberships wouldn't
> be free, a cost of perhaps $500 or so would bring it well within the bounds
> of a book-publishing budget while discouraging frivolity.

Umm.  Why discourage it?  It seems odd to me that the right to 
information which determines the direction of technology and 
technical markets should be sold as if it were a poker ante.  

Don't sell cheaper indulgences.  The W3C should change its rules.

> And once again, I regret that the XML process has failed to meet
> Len Bullard's exquisitely high standards.

Well, by any standards, your reply, Tim, is very civil.  
I respect that and thank you for it. 

len

xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev at ic.ac.uk
Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/
To (un)subscribe, mailto:majordomo at ic.ac.uk the following message;
(un)subscribe xml-dev
To subscribe to the digests, mailto:majordomo at ic.ac.uk the following message;
subscribe xml-dev-digest
List coordinator, Henry Rzepa (mailto:rzepa at ic.ac.uk)