Re: Recent ERB votes

At 11:20 PM 10/11/96 -0600, Len Bullard wrote:

 ... a whole bunch of stuff that deserves a reaction. ...

I know Len quite well and respect his opinion on all sorts of
things, so at this point, I am going to discard diplomacy and say:
Len, you are right out to lunch on this one.

1. Microsoft and VRML and XML

First off, your attempt to draw a parallel to the VRML war stories
is bogus.  I was there too.  In VRML, Microsoft ignored the process,
did not even listen to the discussions, then dropped a violently
incompatible, radically weird, technology from a great height, and
was [deservedly] rejected.  Len also conveniently ignores the fact that
a small panel of self-appointed experts, the VRML architecture group,
once the basic framework was selected, made essentially all the crucial
detail decisions about what was in, what was out, and the
technical details.  Many members of the VAG never said a word on
the public mailing lists.

Microsoft has been on board, since day 1, and has subscribed to the basic
design goals that we pretty much all share.  I would also have been happier 
to  see Jean outline some of his positions to the WG; but if you look at the 
brutal ERB meeting schedule, it becomes clear why only unemployed layabouts
like myself and James Clark, and those who can type faster than 3 people
can talk, such as Michael S-McQ and Eliot Kimber, can really afford to pitch 
in there and also carry lances in WG-land.  The main role of the ERB was to 
listen to the WG, and to vote.  We did those things.

2. Role and Transparency of the ERB

I think that if we could have had a much larger-scale, slower, more
collegial process, we would have had a better XML.  But we would
have had it in 1998, with luck.  The ERB was the only way to make this
happen in Internet time.

Secondly, all the votes of the ERB are a matter of public record.  You
can certainly find some trends there (Bray voting for minimalism at all
costs, Maler for SGML micro-consistency, Kimber for a larger subset,
Sperberg-McQueen for a richer authoring environment) but you can *not*
find evidence of any sleazy corporate machinations.  Because there
weren't any.

Consider the makeup of the 11 voting members.  5 have no vendor
affiliations (Bray, Clark, Kimber, Magliery, Sperberg-McQueen) 3
come from established SGML vendors (DeRose, Maler, Sharpe), 2 from
big system vendors (whose SGML plans if any are opaque to me at
least) (Bosak, Hollander), and Paoli from Microsoft.  If you think that
Jean, even if he had been pushing a diabolical corporate agenda, could
buffalo this collection of obsessive-compulsives, pedantic purists, and
get-it-done-now engineers... gimme a break. 

3. On the nature of appropriate debate

It is perfectly reasonable to disagree with the selection of predefined
attributes (I do) and politico-technical design compromises (as Lee and
Bob Streich do) and deviations from 8879 no matter how minor (as Charles
does).  It is not appropriate or useful, at this point, to start sliming
the process because it producec some results you didn't like.  
This conspiracy soapbox is unworthy of you, Len, and has no basis in fact.
Climb down.

Cheers, Tim Bray
tbray@textuality.com http://www.textuality.com/ +1-604-488-1167