Re: New tags. (fwd) -Reply (fwd)

Jim Wise (jw250@columbia.edu)
Tue, 4 Feb 1997 04:59:33 -0500 (EST)


Date: Tue, 4 Feb 1997 04:59:33 -0500 (EST)
From: Jim Wise <jw250@columbia.edu>
To: Dave Carter <dxc@ast.cam.ac.uk>
cc: Peter Flynn <pflynn@curia.ucc.ie>, www-html@w3.org
Subject: Re: New tags. (fwd) -Reply (fwd)
In-Reply-To: <Pine.GSO.3.94.970204084740.22520E-100000@cass26>
Message-ID: <Pine.SUN.3.95L.970204041314.16273A-100000@ciao.cc.columbia.edu>

On Tue, 4 Feb 1997, Dave Carter wrote:

> Rubbish, 3.2 and Cougar are not standards, the only body which has the
> right to define standards is ISO. W3C certainly doesn't. In the absence
> of ISO activity in this area I suppose the best we have is IETF, so if
> there is any standard it is HTML 2.0 (RFC1866) and tables (RFC1942).

As the old saw goes, `standards are great -- there are so many to choose
from'.  There are many standards bodies, of which the ISO is but one.
The W3C is another, and has for some time been the locus of HTML/HTTP
standards development, with it's `recomendations' being passed on to
the IETF, and used as a general grounds for implementation.  In contrast,
the IETF working group on HTML is specifically chartered to describe
existing practice first and explore new developments only as a secondary
goal.  (See the IETF HTML WG charter [1] for details)

It is the nature of the internet that which standards bodies are
heeded is as much a feature of current opinion and the quality of their
output as of any real precedence among standards bodies.  If you doubt
this, go looking for a site running on a full ISO/OSI network suite...[2]

HTML 3.2 is a process of the W3C recommendation process.  HTML 3.0 is not
(or more accurately, is a reject of this process).

> Personally I find HTML 3.2 and Cougar totally unacceptable, for a start
> they don't include <MATH>. HTML 3.0 is acceptable if incomplete. HTML

But Cougar (which is not a recommendation yet, and has no more real standing
than many other proposals currently floating about) adds the <OBJECT> element,
and cascading style sheets, which hopefully, in their final form, will
provide a much more general-purpose solution to the issues <MATH> was
intended to address.  Due to the rejection of 3.0, features of 3.0 which are
not already supported (does anyone but Arena support <MATH>?) are unlikely
to be added in future browser revisions.  In contrast, HTML 3.2 provides a
feature set which it is safe to _depend_on_.  Similarly, if accepted,
cougar will provide a unified, dependable interface for features such as
stylesheets and embedded objects.

> Pro includes things I would find unacceptable to use, but thats a matter
> of taste, it is a perfectly good DTD, and any process of standardisation
> should start from it.

But HTML Pro is not a standard either.  If I understand (and correct me if I
am wrong), the purpose of HTML pro is to explore possible coherent syntaxes
for a variety of implementation ideas which are floating around out there.
It is very useful in this regard, but I don't think anyone is suggesting that
it is a defintion of what features can be expected in a browser, and it seems
far too much of a moving target to suggest that it is a goal for browser
implementors to shoot for...

====
[1] http://www.ietf.cnri.reston.va.us/proceedings/96mar/charters/html-charter.html

[2] There is also good discussion of this with regard to the current IETF/W3C
    situation in the minutes of last March's HTML WG meeting, at:
    http://www.ietf.cnri.reston.va.us/proceedings/96mar/area.and.wg.reports/app/html/html-minutes-96mar.html

--
				Jim Wise
				jim@santafe.arch.columbia.edu
				http://www.arch.columbia.edu/~jim
				* Finger for PGP public key *