How do we introduce tables into WWW and/or HTML?

"Daniel W. Connolly" <connolly@oclc.org>
Date: Thu, 16 Jun 94 12:53:51 EDT
Message-id: <9406161651.AA02277@ulua.hal.com>
Reply-To: html-ig@oclc.org
Originator: html-ig@oclc.org
Sender: html-ig@oclc.org
Precedence: bulk
From: "Daniel W. Connolly" <connolly@oclc.org>
To: Multiple recipients of list <html-ig@oclc.org>
Subject: How do we introduce tables into WWW and/or HTML?
X-Listprocessor-Version: 6.0c -- ListProcessor by Anastasios Kotsikonas
X-Comment: HTML Implementation Group

I think we made some mistakes in introducing the level 1 and level 2
features. I think it was a mistake to put	

	EM, STRONG, CITE, DFN, KBD....

all in the standard DTD before they were widely deployed and
consistently used. I took the list from TeXinfo, but in TeXinfo, those
tags are used to build indexes and such -- they actually trigger real
functionality beyond just the typographical distinctions (I think).

I think it would have been better to just propose
	EM, BF, and TT
as per TeX/LaTeX, and see how they were used in practice, and
then after a while, revise the spec to describe the working
practices.

I also think it was a mistake to make no distinction in the HTTP
protocol between documents with forms and documents without forms.

With this experience in mind, consider the following two scenarios:

	July 4, 1994:	NCSA releases NCSA Mosaic 3.0, with
			"Full HTML+ support"

			(Exactly what do they mean by HTML+?
			Which DTD are they using? Well... they're
			not using any particular DTD, but they
			support tables and figures!!!)

	July 5, 1994:	Zillions of information providers start
			sticking NCSA Mosaic-style table markup in
			their HTML documents

	July 10, 1994:	Developers of Chimera, Lynx, and all the
			other WWW browsers get tons of email asking
			why they get these funky looking documents.

			"Those documents use table markup, and we
			don't support that yet," they say.

			"When will you?" respond the users.

			"RSN," reply the frenzied developers.

	August 1, 1994:	Lynx announces table support.

	August 2, 1994:	Lynx developers get tons of bug reports
			cuz their tables don't work quite like NCSA
			Mosaic's.

Now consider this alternative:

	July 4, 1994:	NCSA releases NCSA Mosaic 3.0, with
			"Full HTML+ support"

			(What exactly do they mean by HTML+?
			What DTD are they using?
			Well, they took the 2.0 HTML DTD, added
			some stuff from Dave Raggett's HTML+ DTD,
			and tweaked it a little. It's available
			online with the rest of their documentation)

			NCSA Mosaic 3.0 writes
				Accept: text/html; level=3
				Accept: text/html
				...
			in all HTTP transactions.

			Information providers are notified that they
			must use the extension "html3" for documents
			with tables, and server configurations must
			be upgraded to include a mapping:
				*.html3 -> text/html; level=3

	July 6, 1994:	Zillions of information providers make documents
			with tables available.

	July 7, 1994:	NCSA Mosaic 3.0 users successfully browse those
			documents with tables.
			
			Users of NCSA Mosaic 2.4 and other browsers
			get a save-to-disk dialog for documents with
			tables.

	July 8, 1994:	NCSA httpd is enhanced to be able to convert
			level 3 html to level 2 html via an
			external conversion script, and to cache the
			result.

	July 9, 1994:	Some sites enable the html
			down-translator, and some don't. Some sites
			batch convert html3 documents and make both
			versions available through fomrmat negociation.

	July 7, 1994:	Developers of TkWWW, w3-mode.el, Viola
			release patches so HTTP transactions will include:
				Accept: text/html; level=3

	July 10, 1994:	Users of NCSA Mosaic 2.4 and other browsers
			sometimes get plain-text tables (from the converter),
			and sometimes they just get a save-to-disk
			dialog (from the sites that don't make text/html
			available).

			Users of TkWWW, w3-mode.el, Viola browse
			table documents seamlessly.

	July 15, 1994:	HaL releases a batch HTML validation tool that
			scours a whole tree of documents and produces
			a report of the errors, and an
			interactive validation tool that interfaces to
			Mosaic and allows authors to validate their
			documents while they preview them.

	Aug 1, 1994:	First release of the HTML 3.0 spec comes out,
			with tables included in the new level 3
			section of the DTD.
			
	Aug 30, 1994:	HTML 3.0 spec published as a W3O spec. (and/or
			as an SGML Open technical report)


Perhaps there are other scenarios. Perhaps this isn't something
we need to concern ourselves with at this point. But I'm very
curious about how it will happen.

Dan