W3C home > Mailing lists > Public > w3c-wai-gl@w3.org > October to December 1997

Re: TABLES mentioned in style sheet section.

From: Jason White <jasonw@ariel.ucs.unimelb.EDU.AU>
Date: Thu, 11 Dec 1997 10:49:56 +1100 (AEDT)
To: HTML Guidelines Working Group <w3c-wai-gl@w3.org>
Message-ID: <Pine.SUN.3.95.971211101917.18171B-100000@ariel.ucs.unimelb.EDU.AU>
On Wed, 10 Dec 1997, Al Gilman wrote:

> ASG:: Enforcing some "meaning" or "intention" for HTML structures
> such as TABLE that is narrower than the range of useful things
> that the structure can and will do for you will likely be futile.
> 
> HTML is a win because it is so natural.  The rule of natural
> language is "Whatever works."  This is short for the orthodox
> sentence "Whatever works, goes."  The syntactically unorthodox
> form is frequently used because it is usually understood.
> 
> Occam's razor applies to languages and media.  The fewer rules
> one has to follow to get one's point across, the more desirable
> the language or medium.  
> 
JW:: However, it is the semantic value of the HTML markup which is relied
upon by, for instance, braille and audio formatting software to produce an
appropriate rendering in these media. Such a process requires consistency
in the meanings conveyed by the HTML elements and attributes, through
which the author's intention is manifested. Only under this condition
especially in the absence of braille or audio style sheets supplied by the
author, can the output software transform the structure and content of the
document into a format that will be appropriate, without misleading the
reader.

The question which thus falls to be considered, and this is best done with
the aid of examples rather than by attempting a purely a priori analysis,
is whether those structures which, semantically at least, are not genuine
tables would be misunderstood or inappropriately formatted if the
formatting algorithms that a braille or audio application would ordinarily
apply to tables were invoked. To the extent that such a problem would
arise, there is a need for authors to avoid table markup in cases where it
is so contrary to the intended semantics as to result in misleading output
in different media.

Two types of solution are available. Either the bad practices could be
dispensed with in favour of CSS positioning, or those practices could be
accepted, but accommodations made in the HTML markup (E.G. standardised
class names) which alert formatting software to the fact that what appears
to be a table is actually not so. My preference is for the first of these
options. To the extent that the decision of whether to use table markup or
style sheets will often be made automatically by authoring tools rather
than by authors themselves, there will be no additional burdens imposed on
the developers of HTML documents once CSS positioning is widely available.
In the long term therefore, I think the balance of these considerations
leads to the conclusion that the abuse of table markup ought to be
discouraged and that more appropriate strategies, involving style sheets,
provide a semantically appropriate solution that satisfies the
requirements of all media.

Natural language affords a variety of means by which unorthodox usage can
be explained in order to avoid misunderstanding. HTML, is much less
flexible in this respect, and therefore the intended correlation between
markup and meaning, as defined in the specification, needs to be observed
more strictly. The formatting of documents in braille and audio media is a
process which highlights the fundamental importance of this requirement. 
Received on Wednesday, 10 December 1997 18:50:19 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:46:57 GMT