WebAIM - Web Accessibility In Mind

E-mail List Archives

Re: address tag

for

From: Jukka K. Korpela
Date: Feb 15, 2007 5:50AM

On Thu, 15 Feb 2007, Tim Beadle wrote:

> some purists take issue with Microformats on the basis that
> namespaced (X)HTML should be used instead.

That's irrelevant as such. What matters is whether markup is based on
well-designed, publicly reviewed and stable specifications _and_ whether
it becomes widely used. We have a long tradition of great confusion -
consider the rel="..." and rev="..." values for links. Everybody and his
brother invents new values and doesn't even bother defining what they are
really supposed to mean, as a matter of unambiguous semantics.

> Microformats power smarter
> web pages and applications *now*, not in some far-off future.

Such promises mean splitting the Web into incompatible microwebs. Luckily,
it is mostly just idle talk, hype, and pompous plans.

> Equally, adding Microformats to web sites will have no negative impact
> on accessibility, while adding richness for everyone.

Except when the particular microbabble happens to conflict with other
microbabble systems that are actually supported in some software, _or_
with processing of tags by their old semantics or by legacy rules. There
are so few tags and attributes in HTML that you'll end up with conflicting
with _something_ rather soon. Even <span title="...">...</span> isn't
harmless, since the title attribute _may_ be rendered to the user, perhaps
by special request, in a visible or audible way, even if a microformat fan
wanted to use it just for specific purposes to be handled by some fancy
"semantics aware" software.

Regarding the <address> element, it _would_ be quite useful for
accessibility purposes and otherwise, if it were used consistently.
Browsers could have a simple command "show document author's contact
info", for presenting the <address> element's content to the user.
This would be very useful on large complicated pages that might have such
information in some hard-to-find place.

However, the <address> element is more often than not used just as a
container for _any_ address. This has happened despite the fact that
several authoritative published specifications have defined it as an
element containing contact information on the author of the document
(or part of a document). This was partly a misunderstanding caused by a
poorly chosen element _name_ (<address> instead of <author>), but such
misunderstandings _will_ happen for microformats, since the names used in
them will be poorly chosen and poorly defined.

"Semantic web" is a great idea, but it won't even start getting realized
before the current "Semantic Web" fashion has been killed and buried
decently.

The _practical_ way to be semantic on the web at present, and in the near
future (and I mean years and decades), is to express your content verbally
in a clear, unambiguous, and easily understandable way. Words matter.
Search systems are being refined, but mostly so that they process _texts_
(and links between texts) in more elaborated ways.

--
Jukka "Yucca" Korpela, http://www.cs.tut.fi/~jkorpela/