Re: Moving longdesc forward

Benjamin Hawkes-Lewis, Wed, 4 May 2011 15:52:37 +0100:
> On Wed, May 4, 2011 at 2:37 PM, Leif Halvard Silli:

> Why we should constrain @longdesc to resources that are HTML/XML markup
> with an HTML root element?

To say that users plus user agents are likely to expect HTML ( = 
content with <html> as root element) and that they may not be able to 
properly/effectively handle other formats (which may require plug-ins, 
external programs etc), is just that - a warning and a recommendation - 
and not a constraint.

> This constraint would not prevent images, unstructured text, or
> text encoded in unrecognized vocabularies being supplied as text
> equivalents, it just requires them to be contained in a 
> lightweight HTML/XML wrapper.

The format recommendation/warning cannot prevent 'light wrapper' misuse 
- other language in the spec text prevent such misuse, it is supposed 
to be a description! But it can help authors to create descriptions 
that user and user agents can - technically - handle properly and 

Though, as this example shows, longdesc document should probably not be 
inside a framge page:

Benjamin Hawkes-Lewis, Wed, 4 May 2011 15:54:39 +0100:
> On Wed, May 4, 2011 at 2:36 PM, Leif Halvard Silli:
>>> The purpose of @longdesc in the suggestex text is twofold:
>>>     1. Structured text alternatives.
>>>     2. Long text alternatives.
>> Structured *and* (potentially) long.
> Long OR structured.
> If we think @longdesc should only be used where structured markup is
> required, we should revise the requirements and text to match.

'long descriptions' are short pieces - like this one by Laura:

If we agree to recommend HTML (as described above), then we can put the 
long/structured discussion to the side.
Leif Halvard Silli

Received on Wednesday, 4 May 2011 15:49:18 UTC