xml-dev - Anti-Ranti

[Rick JELLIFFE <ricko@allette.com.au>]

> Matthew Gertner wrote:

> I've placed a new rant on our website....
> Any and all comments and (especially) criticism would be most welcome.

You write "There is broad concensus in the XML development community
that W3C must change.." and this is where I think you go astray. Most
XML developers are not on XML-DEV, they don't have time or inclination
to participate, they don't make it to conferences. Perhaps they don't
live in the US anyway...indeed, they might be suspicious of any
"democratic" efforts which replace a far-off institution which can be
pressured with just another mob of loud Americans.  There is no broad
concensus; people don't care because they could nor participate anyway. 
When did we vote; when did you take a poll; how did you track everyone
down?

There is plenty of agitation from people who feel left out or who are
professional journalists interested in whipping up a story (which is not
to say that they are not also developers with a legitimate point of
view, and valued members of the "XML development community", nor that it
is not good to review important issues). Why not? All part of the game.
But strong feelings by some does not constitute a groundswell. Look at
SML-DEV--everyone is bending over backwards to give them time to talk at
conferences, but it is only a handful of excited people: look at SML-DEV
and you see maybe 6 names do 95% of posts.

The number one issue for democratization of the WWW is how to accomodate
input from members of cultures which are based on discretion, deferring
to those you respect even when they are wrong, and politely waiting to
be asked for an opinion at an appropriate time.  

And its not just other cultures. Why are there no women participating in
XML-DEV? The only women whose names I have seen in the XML forums have
either come in from the SGML world (because women are very active at all
levels of the publishing industry) or are involved with W3C.
Occassionally a woman pops her head up and asks a question, but clearly
XML-DEV is not a forum which many women feel inclined to participate on.
If anyone is thinking about opening up standards-making, I would like to
know by what process they will make sure that women are attracted rather
than repelled? As an assertive, Western, white, English-speaking male on
a forum largely made from the same, it seems  strange to hear what seems
to be complaints against W3C/ISO which, however unsatisfactorily,
provide the only forums I have seen which have significant
representation from people who are less assertive, less Western, less
white, less English-speaking and less male than me.

Accomodating outsiders requires

	* some forums should be moderated by someone respected and responsible,
to prevent monopolization by loudmouths, dirty hippies, slick company
men, and any other troulesome stereotypes;

	* decent time intervals between when proposals are made and when they
are decided, to allow consideration by people for whom English is not
their first language (and, please note, that the "legalese" of standards
may be much easier to understand by a technical non-native speaker than
many of the postings on XML-DEV! Indeed, anyone who complains about
legalese should refrain from mentioning any "design patterns" in their
postings--all jargon is mystifying and irritating to outsiders);

	* fairly cohesive point-releases of working drafts, not day-to-day
releases; too little time between releases means that there will be not
enough time for people to formulate an opinion and the releases will not
be choseive, but too much time between releases means that many comments
may be obsolete (given that most obvious problems that external reviewer
finds will also be found by an internal reviewer);

	* reliance on internet technology which does not require people meet at
a particular place or time; 

	* the support of competing specifications: which is better Schema
language?--DTDs from ISO which attempt to be a 80:20 solution but many
people think are too small because of the lack of datatyping; RELAX from
Japan which takes modularization and ad hoc schema composition as the
central problem; DSD from Europe which takes CSS as a good role model;
XML Schemas from US and UK which says that the central issue is
supporting type derivation to allow meaningful local names for standard
types; or Schematron from Taiwan which says all the others are just
minor variations on an uninteresting and hard-to-implement theme (Xerces
Schema support takes over 30000 lines of Java, Schematron takes a couple
of hundred!).

It is this last point which, I think, is the most important. And that is
where, to me, many of the people against W3C fall into the exactly the
same boat as the W3C. 

The issue is whether competition and plurality and diversity is good or
bad. Look at a lot of the rhetoric about SML--it is not just that XML
has been made with certain tradeoffs, they say that XML is positively
wrong-headed for not having been constructed on minimalist principles.
Look at the discussion on "why do we need XSL when CSS or Java is good
enough?", and before that to "XML Schemas should replace DTDs" and
before that to "XML will kill SGML". In all cases, there is a certain
mindset or personality involved which says not "let the market decide"
but "plurality is bad".

The Internet has thrived on supporting plurality: each layer of the
network allows multiple higher layers to declare themselves. URLs start
with a method. HTTP data gives a mime type. XML documents have a
style-sheet PI. We can choose SAX or DOM. We can choose particular SAX
or DOM implementations. 

The way to make W3C open is not to enforce some particular
organizational strategy on them, it is to make sure that all technology
they (and we) create allows 
 * plurality at the next-higher level;
 * does not block out competition at its own level;
 * is specified tightly enough that it can prevent snraling monolopies
from creating incompatible versions--ultimately this can only be done by
a branding and testing program.

In other words, if Tim B-L wants to start up a consortium with enough
clout to get the big players to co-operate more, more power to him.  And
if other people who like the particular technology that results, let
them develop their own: I have been the longest-running and most vocal
critic of the various schema proposals that have come out, but I backed
that up by actually having some technology to contribute rather than
just talk. 

The way to support democratization is to allow encourage plural
development of stadards. There is no reason why, for any problem, there
is one best solution. Certainly there is little reason to have
confidence that we can know it beforehand. Let a thousand flowers
blossom, and a hundred schools of thought contend.  

Don't jump on W3C for being closed, jump on anyone who says "all
technology *should* be constructed along the lines of principle X" or
"everyone should use technolgy Y, so we don't need/want to support
anything else".  (I think Murata-san's comments about RELAX's future are
really good in this regard...oops earthquake...back again...for any who
has read them.)  They are the real conspiracy. 

How do we support respect-based cultures? Respect and allow independent
development and contribution from people in other places and times and
languages and economic circumstances. Clearly distinguish, for each
technology, the mechanism for invoking from the mechanism from
performing the action, and allow the mechanism for invoking to support
plurality.

Tangentially, this is also why I, apparantly alone in the world, think
that PIs are so important--they are the only thing that are utterly
outside the control of the standardizers who dislike plurality (which
often, in a very human fashion, is just that they want to promote their
hard-fought-for technology and they genuinely don't agree with the other
team's technology). No matter how tight someone makes a schema, I can
still put in legitimate markup which allows me to invoke my own
processing, and make that an integral part of the document. Getting rid
of them (without making some replacement, such as a convention that you
can always add elements or attributes from other namespaces and that no
schema language should enforce otherwise--impractical) plays into the
hands of the embrace-and-extenders.

The W3C is not the enemy: perhaps *ML-DEV is!

Rick Jelliffe
Taiwan

(Not speaking on behalf of any employer.)

***************************************************************************
This is xml-dev, the mailing list for XML developers.
To unsubscribe, mailto:majordomo@xml.org&BODY=unsubscribe%20xml-dev
List archives are available at http://xml.org/archives/xml-dev/
***************************************************************************