xml-dev - Re: standards body parallel

[Daniel Veillard <Daniel.Veillard@w3.org>]

On Fri, Oct 13, 2000 at 03:39:23PM -0400, Simon St.Laurent wrote:
> At 09:25 PM 10/13/00 +0200, Daniel Veillard wrote:
> >On Fri, Oct 13, 2000 at 02:45:27PM -0400, Simon St.Laurent wrote:
> >  I'm usually extremely quiet but I will ask you to back up your claim
> >or retire it. 
> 
> For the only public acknowledgment I've seen of such pressure, try:
> http://www.xml.com/axml/notes/JeanPa.html
> 
> To be honest, I'm surprised that those comments were published, though they
> weren't exactly top secret.

  Our process is public:
http://www.w3.org/Consortium/Process/Process-19991111/tr.html#Reports

-------------------------
  6.1 General information about documents 

  Each document produced by a group will be edited by one or more
  editors appointed by the group Chair. 
-------------------------

  Editor appointment is not a vote, it was the responsability of the
chair (Jon Bosak in this case) to handle the objection. The chair had
to handle a case where someone objects to an editor suggestion. In the
end a compromise was found, but strictly speaking it's not about the
decision of the content of the specification, nor it's accession to
REC status. I think it was purely a PR issue as Tim seems to indicate.
Editors don't have power over the content of the specification, though
in practice if you are the one editing a document it is usually easier
to suggest modifications to the Working Group.

> For those of us not at the table, a vote is not a vote.

  That's another issue. We have invited expert status for those who can
contribute significantly while not part of a member organization.

> While I would love
> to take your statements at face value, there is no way to verify such
> claims without opening the archives.  My understanding is that a vote is a
> vote is a vote, except in certain rare situations.  

  And the example you provided is not one subject to vote.

> exactly Len Bullard is saying.  I think you may be deeply irritated that
> I'm questioning your views, but I'm hardly attempting to minimize them.

  Okay, you have been carrying a message saying more or less that
centralized design done at W3C where not everybody is invited to the
discussion is not a good model. And also suggesting, but in relatively
vague terms that a completely decentralized open design would be a 
better choice for XML standards ...

  You promoted the SAX example as one of the case where this succeeded.
I would say that personally I have troubles with SAX in the sense that
nobody took the time to see how it should be done in pure C, though one
of the most successful XML implementation expat uses one close to it but
there is no "official" mapping for C. So for me at least this example 
is not perfect. Nobody in this decentralized open assembly decided it
was worth its time doing it for C, and I think that's the crucial problem.

  Doing the 80% of the work takes 20% of the time what is painful is getting
the 20% left right and the 80% energy it requires to get it done. Don't
get me wrong, I'm a Linux and free software advocate from far before I
joined W3C and I'm involved in the Gnome project where I donated the
time needed to build the libxml library. But I have also seen the 20%
problem there too ... It's hard when there is an assembly of free peer
to get someone dedicated to do this 20% of the dirty job which is the
final touch requiring so much efforts. This is the case of debugging,
documenting, preparing releases, packaging that were lagging behind in
the free software projects I was a member of before some benevolent
*paying* firm bought the dedication of some people to do it. It was
the case until the first commercial releases of Linux appeared, it was
the case of the Gnome project until Redhat and later other startups
paid people to do this dirty job. Now the majority of the people enjoy
having the work finished, but there will always be voices saying that
the paying firm took control other the free process.

  And I think that for pure specifications the problem is even more
difficult. Other factors are scaling, and resolution of conflicts. Both
are easier when there is some kind of control over the resources used 
to attain the goal.

  Last but not least, I do understand how a non-open process may
frustrate people not invited to join. However at W3C it is possible
to get involved through the Invited Expert status. If someone want and
it's clear that you can provide valuable input, ask the chairs to join.
But you will have to agree to the WG rules including the fact that
a minimum garanteed effort has to be provided by the group member
(and is applied indifferently to invited experts and people who
joined from W3C members).

  IMHO suggesting to get everything open and risking anybody to block
the process by criticizing without contributing is actually the worse
model. Even in projects like Gnome not everybody can commit to the CVS
database. And not everybody with CVS write access is entitled to change
someone else code without notice. At the end who does the code get the
final word,

    "Just Do It !"

Daniel

Speaking for himself !

-- 
Daniel.Veillard@w3.org | W3C, INRIA Rhone-Alpes  | Today's Bookmarks :
Tel : +33 476 615 257  | 655, avenue de l'Europe | Linux XML libxml WWW
Fax : +33 476 615 207  | 38330 Montbonnot FRANCE | Gnome rpm2html rpmfind
 http://www.w3.org/People/all#veillard%40w3.org  | RPM badminton Kaffe