E-mail List Archives

Re: abbreviations


From: Jukka K. Korpela
Date: May 28, 2010 12:27PM

Denis Boudreau wrote:

> One option could be that when a screen reader (or any type of
> assistive technology for that matter) parses a page and finds either
> an <acronym title=""> or <abbr title=""> tag, it would store it in
> some kind of buffer memory (possibly flushed out one page reload or
> whatever, but temporarily stocked for reference if need be).

I guess that would be in accordance with some underlying ideas behing W3C
WAI recommendations. This does not necessarily make it a good idea, though.

In practice, the <acronym> and <abbr> tags, with the title attribute, may
have undesirable side effects on graphic browsers.

At the logical level, if it's document-wide information about some
abbreviation, it sounds like a candidate for metadata, rather than
element-specific data. It's illogical to say something about a specific
instance of an expression when you really mean all occurrences of it in a

Resolving the potential ambiguity of, say, the string "IT" (is it an
abbreviation for "information technology", or the pronoun "it" in uppercase,
or something else?) is but a small part of disambiguation, and I would not
give it any particularly important role. Abbreviations can be deceptive, but
human beings can deceive each other and themselves with mere words, too.

Basically, I do not believe in any particularly "high-tech" approach here,
and I count even HTML markup, the poor lonesome data format invented in the
early 1990s, as "high-tech" as compared with normal written or spoken text.
When you speak or when you write, say, e-mail or SMS or a book, you don't
have the questionable luxury of using markup for (purportedly) resolving
ambiguities that may arise in your text. You just have to be careful, and
that's what I recommend to web authoring, too. Try to make your context, or
sometimes explicit explanations, disambiguate when needed.

(Disambiguation of words and abbreviations is just one part of the story.
Quite often, expressions are grammatically ambiguous, i.e. can be parsed in
different ways that result in different semantics. Markup is hardly a

Yucca, http://www.cs.tut.fi/~jkorpela/