Re: Longdesc change proposal update

On Sun, May 8, 2011 at 12:56 PM, Silvia Pfeiffer
<silviapfeiffer1@gmail.com> wrote:
> Hmm, that text could do with a bit of technical clean-up. May I
> suggest some revised wording? How about something like the following?
>
> ===
> The longdesc attribute may be present and contains a valid non-empty

I think it's better to use formulation "must be" rather than "contains",
to be clear we're talking about an authoring conformance requirement.

cf. @cite at http://dev.w3.org/html5/spec/grouping-content.html#the-blockquote-element

> URL potentially surrounded by spaces referencing a Web resource that
> contains detailed information about the image that the Web page author
> wants to make available, but not in the main flow of the Web page.

A text alternative is not primarily "detailed information about an image",
but text serving an equivalent purpose.

In particular, we want to avoid @longdesc being misused (as on Wikipedia) to
refer to a page of metadata about the image that does not include text serving
an equivalent purpose.

This is why our draft text follows WCAG2 in using the phrase "text alternative":

http://www.w3.org/TR/WCAG20/#text-equiv

> The longdesc resource will in particular contain a textual description of the
> content of the image

"A textual description of the content of the image" is not necessarily
text serving an equivalent purpose.

> that accessibility technology will make available to vision-impaired users.

In line with WCAG2, we should present vision-impaired users as one of many
groups who can use text alternatives, not the only one:

http://www.w3.org/TR/UNDERSTANDING-WCAG20/text-equiv.html

The usual expansion of "AT" is "assistive technology"; this is the phrase
ARIA uses for example.

I'm not sure what "will" is doing here. Is it supposed to be a conformance
requirement of some sort? AT is not currently a distinct HTML5
conformance class.

> The user agent should expose the longdesc link to the user, but it
> should not interfere with the page's normal rendering.

This formulation does not seem compatible within inline replacement,
or with user agents opting out of the "normal rendering" part of HTML5.

> For example, it could be exposed in the image element's context menu.

This sort of implementation suggestion is best reserved for the
rendering section, as with @cite.

Note the rendering text we've drafted:

http://www.d.umn.edu/~lcarlson/research/ld-rendering.html

> Note: The longdesc page can be regarded as the information page of the
> image. It should in particular contain a detailed description of the
> content of the image such that vision-impaired users can also
> understand what is presented in the image. It may contain structured
> markup, such as a table to explain a complex graphic like a statistics
> chart. It can also contain the image itself, links back to all the
> pages that contain that image, and metadata about the image such as
> license and copyright information.

This note does not make it clear that in this case the @longdesc attribute
should point to the "detailed description" fragment not the surrounding
paraphernalia.

--
Benjamin Hawkes-Lewis

Received on Sunday, 8 May 2011 12:29:56 UTC