Re: HTML3 Tables

lee@sq.com
Wed, 8 Mar 1995 12:07:40 -0500


Date: 	Wed, 8 Mar 1995 12:07:40 -0500
From: lee@sq.com
Message-Id: <9503081707.AA08475@sqrex.sq.com>
To: Jon_Bosak@novell.com
Subject: Re: HTML3 Tables
Cc: www-html@www10.w3.org

> > We (SoftQuad) have a content model for this that's in the public domain,
> > and has also been published as an ISO standard.

> Literally a superset?  Will all tables valid according to the current
> HTML 3.0 table model be valid according to the SoftQuad table model?
Well, I did say that it doesn't have the border attribute right now, so
strictly speaking no.

We use architectural forms to change the element names in our table
fragment, so for HoTMetaL we are using the
	(TH*,TD*)
model (from memory, I think that's right) for a row.  I can't say whether
all HTML 3 tables are valid, as I haven't sat down with the latest 3.0
draft -- but what I claimed was that it was a superset of the NCSA Mosaic
tables, except without any border element.  I think it's pretty close,
though.

The major difference is that HoTMetaL uses COLDEF elements if they are there,
to specify the default formatting attributes for each column of the table;
there's one empty element for each column, all in a COLDEFS container for
convenience.  HoTMetaL always writes these out, but it doesn't require
them on import.  The COLDEFS section lets you specify relative column
widths, for example, which is pretty useful.

> Which ISO standard has it been published in?

(Ulp) I'll have to check, I'm no good with numbers, it got in with the
annex to the ICADD report, though.  Jeff or Yuri could confirm that and
supply the number, or can say I'm wrong and it didn't make it :-).

Note:
There are actually some differences between ICADD and SQ tables, but as
I recall the differences are unlikely to affect HTML -- it's mostly to
do with the way you do tables with a running head and foot as 3 separate
tables, and then expect the column widths to match up...

Lee