Re: Initial Draft --Cascaded Speech Style Sheets

David Seibert (
Wed, 14 Feb 1996 17:25:36 -0500 (EST)

On Wed, 14 Feb 1996, Scott E. Preece wrote:

> From: David Seibert <>
> | ... The best way to achieve this unity is
> | by making authors use descriptive rather than prescriptive tags, e.g.,
> | <strong> rather than <b>, which can be rendered in whatever appropriate
> | aural or visual styles are available to the UA. This requires a change
> | in html, rather than in the style sheets, but it is not a difficult
> | change. The simplest way to achieve this is to make the prescriptive
> | tags redundant, giving them the value of the nearest equivalent
> | descriptive tag, which would preserve the backward compatibility of html
> | documents.
> ---
> I agree with the idea of promoting descriptive markup; however, the key
> to getting better markup is to give authors the choice of more
> descriptive tags.

Good point!

> While <EM> and <STRONG> do remove the implication of a specific visual
> style, there is NO rationale for using one or the other for any given
> piece of text other than imprecations (where the strength of the warning
> is intended). While their use for warnings *does* imply a way to
> distinguish them aurally, their use for anything else does not. Aside
> from that one use (which would be better conveyed by WARNING and
> STRONG-WARNING tags), they are just presentation markup, but without the
> virtue of convention to support them.

The tags <em> and <strong> have more intuitive meanings than <i> and <b>
for people who work aurally rather than visually. Their usefulness is
not limited to warnings - think about how you would say "<b>Stop</b>
pressing the button", and "If this were a <em>real</em> emergency".

> If, in writing bibliographic citations, I use CITE for the title, STRONG
> for the volume number, and EM for the issue number, that just means I
> want them distinguished, it doesn't mean I want the volumne number
> shouted. There are typographical conventions that make it appropriate
> to use I and B for specific kinds of data; while it would be better to
> have descriptive tags for those kinds of data, EM and STRONG aren't
> those tags. There is no way it makes sense to code a foreign word
> (which would conventionally be set in italics) as EM.

That's a good argument for keeping the prescriptive tags for visual
presentation when they don't carry any meaning, but suggesting that they
be ignored by aural UAs as the default. That way, authors will be
encouraged to use descriptive tags to get better multiple media

> Achieving descriptive markup requires adding a bunch of new,
> descriptive tags - things like PERSONAL-NAME, DATE, VOLUME, ISSUE,
> CHEMICAL-COMPOUND, etc. When you suggest that people hold their fingers
> in the sign of the cross and run off muttering about tag explosion.

As you pointed out later, the class attribute has emerged as a way of
sweeping the tag explosion problem under the rug. The idea of using
classes to convey descriptive information is fine with me. I don't think
that teaching it to authors will be the problem, that's what wysiwyg html
editors are for.


Work: Home: 6420 36th Ave.
Physics Department, McGill University Montreal, PQ, H1T 2Z5
3600 Univ. St., Mtl., PQ, H3A 2T8, Canada Canada
(514) 398-6496; FAX: (514) 398-3733 (514) 255-5965