xml-dev - WG's responding to XML-DEV (was Re: What is wrong with SVG?)

["Michael Champion" <mike.champion@softwareag-usa.com>]

----- Original Message -----
From: "Peter Murray-Rust" <peter@ursus.demon.co.uk>
To: "'XML-DEV (E-mail)'" <xml-dev@xml.org>
Sent: Sunday, March 05, 2000 6:23 PM
Subject: RE: What is wrong with SVG?


>
> In the present case there is probably no real point in discussing the
> details of SVG *on XML-DEV*. WGs will not normally take notice, unless
> there is so much clamour in the commons that it seems politic. There may
be
> a point in discussing some issues that are generic or cross more than one
> spec. So if a number of W3C specs had attribute values or content that
> require microparsing it could be useful to raise it here. It is unlikely
> that there would be a direct response.

IMHO it is very unfortunate that most WG's take little public notice of
XML-DEV comments until there is "much clamour in the commons".  FWIW,  the
DOM WG kept rather close watch of the public mailing list and XML-DEV when
the details of the Level 1 spec were being hammered out two years ago, and
continues to give very quick and detailed responses on the public list for
Level 2.  I believe that our public debates (I especially remember tangling
with Don Park and Steven Savitsky) helped clarify things greatly.  For
example, while I was mad at the time that that we dropped the NodeIterator
interface from Level 1 because of all the difficulty in explaining it to the
public, it's clear in retrospect that this needed much more "cooking" to get
right.

The counter-argument seems to be that WG members  have little time for this
task, and a more optimal use of limited resources is to simply consider
public comments behind closed doors.  (I must confess that for various
reasons I had nearly full-time to devote to the DOM during the period I was
editing the Level 1 DOM spec, so I *did* have spare bandwidth to devote to
XML-DEV). Nevertheless I believe that W3C participants *have* to find the
time, or cut back their objectives to make the time.  For one thing, people
are discouraged from writing closely reasoned critiques if they fall down a
black hole until the "disposition of comments" when the spec is finalized
months later.  A genuine debate motivates everyone to have their facts and
logic straight rather than just wielding the flamethrower. Also, given the
relative costs of fixing software problems in the field vs fixing them at
design time, the W3C and its members should have a very strong incentive to
encourage detailed critiques before Last Call, by which time only
"showstopper" problems have a prayer of being fixed.

I don't follow SVG closely, but I would encourage people to debate the
"subelements vs microparsing" issue publicly, marshall whatever evidence can
be obtained on each side, and make SURE that the WG responds if it appears
that they are on the wrong track here.  SVG is far too critical to the Web
to allow it to be suboptimal.  Likewise if the inefficiency of the DOM is
part of the problem here, that should be addressed  (perhaps with a hybrid
push/pull model or a "flyweight" mini-Node), not force  people to kludge
around it. The only practical way to figure out whether the SVG and/or DOM
specs need to be improved is for *lots* of people to analyze and critique
them, and the XML-DEV subscribers are probably the only semi-organized group
in the world qualified to do so.



***************************************************************************
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/
***************************************************************************