W3C home > Mailing lists > Public > www-html@w3.org > December 2003

Re: Abbreviations and Acronyms: [techs] Latest HTML Techniques Draft

From: Jukka K. Korpela <jkorpela@cs.tut.fi>
Date: Sat, 13 Dec 2003 01:46:11 +0200 (EET)
To: w3c-wai-ig@w3.org
Cc: www-html@w3.org
Message-ID: <Pine.GSO.4.58.0312130118420.17581@korppi.cs.tut.fi>

On Sat, 13 Dec 2003, Charles McCathieNevile wrote:

> Sidar's WCAG2-espa group has discussed this a bit, and the emerging
> consensus basically agrees with what Christian says below (which saves
> me a lot of typing ;-).

Sorry, but that sounds very odd. But first I'd like to ask for
clarification for this:

> The point about expanding all instances, not just the first, is
> important to work with current technology - a paper presented by Sofia
> Celic looked into this. It's the kind of thing tools should do anyway -
> it's a pretty simple search/replace type script.

Does that mean that the _markup_ should have abbreviation markup with a
title attribute for each occurrence of each abbreviation? That would be
strange. Should authoring tools generate such stuff, instead of browsers
simply storing the information found in the first occureences and using
it throughout?

Let's see Christian's points then:

> > - - for <abbr title="for example">e.g.</abbr> - -

This seems to say, rather implicitly, that conventional notations are to
be treated as abbreviations, too. Where will that end? The notation
"e.g." is not abbreviation of "for example" (though it's an abbreviation
of some Latin words, historically - should all such cases be marked up
too?).

> > you'd
> > need at least three classes to be implemented in an aural stylesheet:
> > * spell out element content
> > * read title
> > * read element content

To repeat the question I asked before: is the difference between the
written and spoken forms of an expression really presentational?
And has the title attribute been silently redefined as specifying
the spoken form? I remember the time when it was defined as an advisory
title.

> > So differing between those two kinds of abbreviations that are
> > acronyms and that aren't isn't that important at all, I think.

So are we supposed to use <abbr> markup for all words that are originally
abbreviations? Why?

> > I think the WAI HTML Techniques Draft should state that it's important
> > to
> > markup abbreviations at all, but it's not so important to markup those
> > special abbreviations that are acronyms as such.

I have never seen any good arguments in favor of such a recommendation,
which more or less coincides with what WCAG 1.0 says - though the WAI
pages themselves don't even try to apply those principles. The page
http://www.w3.org/WAI/ has no <abbr> markup, not even for "WAI" itself,
and it has just a few casual occurrences of <acronym>.

> > Also I suggest that abbreviations are always marked up, not just the
> > first time, maybe the title can be given only the first time.

Really? Does anyone seriously expect authors to do so?

What's the big idea in using any abbreviation markup? The main argument,
if not the only one, is related to speech synthesis. But normal speech
simply reads abbreviations either letter by letter or as words, and either
way they are understandable if the text as a whole is. There are some
common abbreviations like "Mr." that are expanded, but it should be very
easy to make a speech generator deal with them - it would be absurd to
require that authors use markup for such very common expressions.

There are a few real problems, though usually not serious problems, with
abbreviations (or notations) that could be read either as words or
letter by letter, such as "WAI" or "ISO". It might occasionally be useful,
thus, to indicate in a style sheet which way is preferred.

It would be generally counter-productive to use abbreviation markup for
all abbreviations _especially if_ user agents started taking it seriously.
Imagine, say, a document mentioning "WCAG 2.0" on almost every line,
read so that each and every occurrence of the expression is expanded to
the full wording.

-- 
Jukka "Yucca" Korpela, http://www.cs.tut.fi/~jkorpela/
Received on Friday, 12 December 2003 18:46:13 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 27 March 2012 18:15:59 GMT