W3C home > Mailing lists > Public > w3c-sgml-wg@w3.org > November 1996

Re: Recent ERB votes

From: Len Bullard <cbullard@HiWAAY.net>
Date: Mon, 11 Nov 1996 17:43:47 -0600
Message-ID: <3287BA33.6F80@HiWAAY.net>
To: "W. Eliot Kimber" <eliot@isogen.com>
CC: W3C-SGML-WG@w3.org
W. Eliot Kimber wrote:
> 
> I guess.  I just didn't want anyone to get the idea that Jean was somehow
> working to subvert the XML process to somehow reflect MS's presumed
> nefarious ends.  

I haven't gone over the edge yet.  We squashed all of the "Microsoft 
is The Evil Empire" stuff on the VRML list and I am the one that stood 
up for them and got the hate mail from the kids as well as some nice
thank yous 
from the MS employees.  I am also the one that took them to task for 
not respecting the community they asked to be part of.  All in a days
work.

But anytime we say HTML now, we are automatically saying Netscape or
Microsoft.  
It isn't even a contest elsewhere.  Furthermore, ecotonal reality: the 
W3C HTML spec may not recognize frames, but the ISO draft for VRML 2.0 
does.  Life at the interface.

That is a market reality for the moment, and because they have the HTML
browsers 
any of us who expect to sell in the HTML market will bend.
That is also a market reality.  Yet, when we begin to grandfather *any* 
existing SGML application, we are on bad ground and will be held 
to it.  SGML doesn't work that way, and if XML does, I think it 
is a disaster. 

That said, do any of you seriously think for a moment we WON'T provide 
HTML tag scanners?  Get real.  Truth is, if it says HTML in the content, 
we won't handle it as SGML.  The Dueling Dwarves have made that
impossible 
and the W3C can't fix the problem.  So, the tag salad is a permanent 
part of the code now.  Burrrrrp!

> You are certainly entitled to hold Jean and MS to whatever
> standard you wish and as a former employee of the largest (and, before
> Microsoft, the most rapacious) vendor in the computer business I can
> sympathize with holding MS to a higher standard in general (whether it's
> fair to hold Jean to that same standard, I don't know).

In this discusssion, Jean is Microsoft.  I will never blame him 
for any move the company makes, but just as I have to stand up 
and take blows for Lockheed's local implementations of SGML, 
he must for MS.  That is how it works.  OTH, I have also had 
to stand before the SGML groups and say, "yes, I know, but 
the programmers just won't do it".  Got the T-shirt.
 
> I can tell you that in the ERB conference calls and e-mail discussions,
> Jean never said anything like "it is MS's official position that" or "I've
> been told by my superiors" or anything.  His arguments were always
> technical, based on his knowledge both of SGML and what he felt he could
> get MS engineers to accept and implement (which is likely different from
> what MS official policy on what they *should* accept and implement might
> be).  

Yep.  And he will argue for that.  No problem.  I would much rather 
see those arguments on this list.  I would much rather see all of the 
ERB arguments on this list.  I dislike oligarchy, especially the 
self-ordained type.  It is political reality, but I repeat, if 
you want support, we have to see the process work:  up front.  

Process makes community.  Nothing else will work.  The only reason the
VRML 
community stood up to assaults was *everyone* voted.  Everyone had 
a stake.  That works.  This process will always be suspicious 
regardless of circumstance.  The W3C is a weak consortium 
and getting weaker by the day.  If you want to make it stronger, 
make it open.  By the time VRML 2.0 came around, wiser heads 
took it to ISO.  I suspect XML has a similar future, but by that 
time, it better be SGML 9x.

> As XML is largely a marketing exercise, we took his statements about
> implementability seriously (at least I did), even if we didn't always find
> them convincing or compelling.  

Marketing exercise?  Call me skittish, but 
the best way to get this to be accepted is if both communities 
see it as an advantage.  If the SGML tool makers and vendors don't,
it will not fly.  That is 98 percent certain.  We already have work 
in progress to put SGML on the Internet.  If XML is a competitor to 
that, or is perceived to be, it will be a major problem here.  

For example, if Dr Goldfarb announces at the end of this process, "nice
try but 
not SGML", XML will have to go elsewhere for support.  I don't 
speak for him, and he certainly will rebut what I say if I 
misspeak, but his position makes him responsible for an 
international standard and to an international body 
supported by nations, not corporations.  You and I are 
aware that he takes that PD seriously. 

> There were certainly disagreements about
> what was and wasn't easily implementable.  But Jean never said (and doesn't
> have the authority to say, as far as I know) "MS will never implement XML
> if it doesn't have X", at least not in a way that could be construed as a
> statement of policy [we all made extreme statements in the heat of argument
> and a lot of them involved the word "never"].

That is the nature of committee work.  Again, no disrespect to Jean, but 
until his arguments appear on this list, I really don't take his
arguments 
seriously because I don't know what they are.  If that is something I 
don't have to know, then fine.  Cooperation isn't a begging exercise.
If our engineers can implement it and his can't, so much the better 
for our market position.

That said, I'm on his side about ease, but ease of use, not
implementation.
We pay programmers to do what they are told.  Since we hire good ones 
and treat them professionally, we expect professional results. 
Microsoft's
personnel policies are their own problem.  But sure, as I said, I know 
about cat herding.

> Note also that in the discussion of grandfathering HTML, the focus was on
> legacy data, not legacy browsers.  

When there are only two significant HTML browsers, that is the same
thing.  
Besides, there is a lot of SGML legacy out there.  If XML has any other
life, why not grandfather that?  The reason given was "political
suicide".  I repeat:  what shotgun compels that?  If it is just <e> vs
</e>, use <e> and ALL of the applications are 
happy.  But to grandfather one app for it's legacy and leave the rest, 
the ERB is on bad ground.

> The question was always *what would the
> marketing effect of making things really different for HTML authors be",
> not, how would this hurt Netscape or IE.  

What about the rest of the world's SGML authors?  Fish or cut bait.

> I think we presumed that Netscape
> and MS are both more than capable of handling whatever syntax we came up with.

I hope so.  I have my doubts.  We have done a lot of work in this area
and 
some aspects of one of those browsers make us very suspicious of its 
extant code base. 

As for our work, we will use SGML.  If XML works there, good.  If not, 
let the market pick products based on criteria they understand. We have
HTML apps ready.  It is the SGML that is work and it is close.  DSSSL is 
next year's challenge.

Dang.  back on the airplane tomorrow....

cheers,

len
Received on Monday, 11 November 1996 18:43:29 EST

This archive was generated by hypermail pre-2.1.9 : Wednesday, 24 September 2003 10:03:42 EDT