W3C home > Mailing lists > Public > public-html-a11y@w3.org > March 2012

Re: aria-describedat

From: Benjamin Hawkes-Lewis <bhawkeslewis@googlemail.com>
Date: Wed, 21 Mar 2012 23:33:08 +0000
Message-ID: <CAEhSh3c2fk+2OEXAdcM8+_RY8NH4zytJQPTajMfPM2FQtPC8oQ@mail.gmail.com>
To: Silvia Pfeiffer <silviapfeiffer1@gmail.com>
Cc: Richard Schwerdtfeger <schwer@us.ibm.com>, W3C WAI-XTECH <wai-xtech@w3.org>, w3c-wai-pf@w3.org, public-html-a11y@w3.org, laura.lee.carlson@gmail.com, George Kerscher <kerscher@montana.com>, david.bolter@gmail.com, jbrewer@w3.org, faulkner.steve@gmail.com, mike@w3.org
On Wed, Mar 21, 2012 at 11:18 PM, Silvia Pfeiffer
<silviapfeiffer1@gmail.com> wrote:
>> - One of the problems with longdesc has been that its data type is not
>> obvious. The name should reflect the fact that it is a URL. e.g.
>> "aria-descriptionurl" or "aria-describeaturl" would be better.
>> "aria-descriptionurl" has the advantage of being a little harder to
>> confuse with "aria-describedby". It surprises me that after all the
>> problems with the naming of "longdesc" people are proposing adding a
>> similarly ambiguous name.
> Interesting point. How would descriptionhref or descriptionref work?

"descriptionhref" is at least similar to "href" which people are used
to using with ordinary links.

"descriptionref" would be bad because it's just as ambiguous (URLs are
not the only type of "ref").

>> - What's the rationale for adding yet more attributes with the "aria-" prefix?
> It's part of the aria set of technologies and developed as part of
> these technologies in WAI, so logically added with the same custom
> prefix as the other aria technologies.

What appears logical to spec writers may prove less digestible to authors …

A prefix that clearly distinguished annotations aimed at accessibility
APIs from markup aimed at general semantics might be useful, but ARIA
has too much of a permanent identity crisis for "aria-" to function as
that prefix.

Compatibility with legacy content is a rationale for preserving the
"aria-" prefix on existing attributes.

Allowing host languages to use the name "describedat" might be a
rationale for a prefix, but ARIA has already reserved the more generic
name "role" without a prefix. A language that featured both
"describedat" and "aria-describedat" would seem especially confusing
for authors.

> We could, instead, propose it as a HTML attribute, drop the "aria-"
> part and make sure it gets into HTML5 and developed by the HTML WG. Is
> that what you're suggesting?

I'm not convinced about the value of expanding ARIA, but that wasn't
particularly what I was getting at. I was just questioning why, if we
expand ARIA, it shouldn't reserve "descriptionurl" just as it reserves

Benjamin Hawkes-Lewis
Received on Wednesday, 21 March 2012 23:33:57 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 15:05:27 UTC