I think you may have missed a trick or two on the way. Dissolving the
WG means moving the focus _away_ from SGML towards some form of
presentation-oriented markup that can make money quickly for a small
number of people (both very desirable goals, just not what HTML was
intended for).
> I assume this silence means everyone assents to it?
It must have happened while I was on vacation. I don't see a need for it,
but I'm happy to be persuaded otherwise.
> Current practice HTML as it stands right now *cannot* be integrated
> into SGML. The reasons for this can (and have been) be argued for years.
> But it is the reality. There are now literally MILLIONS of web pages
> that it is impossible to write a working DTD for.
I have never argued that we should coerce HTML into modeling existing
unparseable pages: that is mere foolishness. Trying to build DTDs for
every kookie idea that browser-writers or authors come up with is a
waste of time: we should critically accept good ideas and forcefully
reject rubbish, and we should be trying to explain to them how they can
get often the same or better results using existing or superior models.
A lot of my (private, paid) time is spent doing up visually-optimized
pages into HTML3 which end up looking better and performing better for
being done right. I'm very happy to let people give me money to do
this, but it's painful to see the look on their faces when they realise
they've been (often unwittingly) duped by the original `designer' into
signing off on inadequate or badly-made material.
> So I repeat: It is time to officially disband the HTML WG.
Not quite yet. We're trying to make a standard which will let people
create durable and reusable information bases, a standard which can
also be used (and misused) for other, more immediate, more ephemeral,
information.
The fact that most (not all) browser makers neither understand nor
care about SGML, should not stop us trying. We should be trying to
persuade tham that their best _long-term_ interest lies in doing it
right, not wrong.
Can we please spend less time arguing about the minutiae and more on
getting the tagset defined and the content models written. I suggested
adding #PCDATA to the content model for TD and TH. I had one comment.
///Peter