RE: ISSUE-194: full-transcript - Chairs Solicit Alternate Proposals or Counter-Proposals

I am not sure that these are necessarily the same thing at all. A transcript is IMO a static untimed merged representation of the information in in the caption and description tracks. A longdesc would probably be something more along the lines of a synopsis or précis. I think we need mechanisms that can handle both of these use cases.

I agree that it makes sense to wait and see how the discussion on generic 'off page text' pans out; it might be for example that we end up with both an attribute and an element e.g. @longdesc and <longdesc> (following the precedent of @src and <source>) where the latter admits a richer set of adornments, possibly including some sort of role attribute which can distinguish between a transcript and a synopsis, amongst other uses for off-page text.


-----Original Message-----
From: Silvia Pfeiffer [] 
Sent: 15 March 2012 03:05
To: Sam Ruby
Subject: Re: ISSUE-194: full-transcript - Chairs Solicit Alternate Proposals or Counter-Proposals

On Thu, Mar 15, 2012 at 12:28 PM, Sam Ruby <> wrote:
> On 03/14/2012 08:27 PM, Silvia Pfeiffer wrote:
>> Dear Chairs,
>> May I suggest that the resolution of this issue is dependent on the
>> resolution of several other issues and discussions currently under
>> way. I therefore propose to delay dealing with this issue until such
>> other issues are resolved. This includes in particular the @longdesc
>> discussion.
>> In my mind, it is possible that the HTML WG decides that there is
>> sufficient need for a general mechanism to add off-page textual
>> representations to certain complex elements to HTML5, elements such as
>> canvas, img, video, audio, table, or figure. If such a general
>> mechanism were added - which could be called @longdesc or @href or
>> @transcript - such a mechanism would fulfill the needs of this issue.
> Can you cite any existing proposal that, if adopted, would address this
> need?

The need for a generic attribute to provide linked lengthy text
representations for complex elements has been emerging.

John's change proposal alludes to this need when referring to
@longdesc [1] as a solution for <video>/<audio> transcript links.

The proposed @describedat attribute which is mentioned in Laura's
longdesc CP [2], and discussed at WAI and in the accessibility TF
points to the same need, ePub even having pressed forward with a

There is a similar need for other elements, too, e.g. <canvas>[4],
<table>[5], <object>/<embed> [6]. I'm sure if we take the time to
analyse this properly, we'll come up with other complex elements that
could be enhanced with such linked transcription resources, too.

There isn't currently a proposal for such an attribute. One reason is
that we've mainly looked at solving these problems with each element
individually and haven't really abstracted this as a common problem
yet. Another reason is that @longdesc could be expanded for such other
elements, but right now it's not even conformant. Basically, we've not
had this discussion yet. This is why I am asking for more time on this
issue for <video>.


[1] "Should
[the longdesc issue] be resolved by reinstating @longdesc in HTML5,
the use of @longdesc to address the use-cases envisioned would likely
be sufficient, making this proposal redundant."
"Protocols and Formats Working Group (PFWG) has discussed introducing
a new ARIA attribute - aria-describedat - as a possible addition to a
future ARIA specification (ARIA 1.1) for external descriptions."
[4] "For animated
graphics demos, a description of the demo could be enough."
[5] - has an example
with a off-page link that is positioned next to the table in a
visually non-intuitive manner
[6] - exemplifies the need for
long text alternatives for flash, which is used through object/embed

Received on Thursday, 15 March 2012 10:02:40 UTC