DC.language Re: January 14 Minutes

Hi Phill, this seems like an issue that belongs in another WAI group, but I
can't think of which one. If I do I'll send them a pointer to this thread.

As I see it, Dublin Core are trying to categorise any resource available on
the Web, and less and less of those will be HTML. So the fact that a Dublin
Core statement duplicates an HTML attribute doesn't seem to be a problem - in
fact if the information is in the HTML it should be easy to extract it for
making accurate (or at least consistent) Dublin Core metadata, or take it
from the metadata and include it in the page encoding. The same applies to
the potential overlap between either of these and information provided by an
HTTP header.

So I propose a Technique for tools to meet WCAG checkpoints 13.2 and 4.3, for
ATAG checkpoints 1.3 and 3.2:

(applicable to all except programming tools, I think. Images don't often have
a language, but audio often does...)

If identifaction of the language of a document is available from any of
 + The document's markup (eg xml:lang attribute in XHTML, SVG, SMIL and
      others)
 + The HTTP headers of a document obtained from the Web
 + External metadata about the document, such as Dublin Core information

propose the value to the value to the author (suggested, but necessary if
there are conflicting values obtained) and use it for all of the above which
are applicable.

cheers

Chaals

On Mon, 14 Jan 2002, Phill Jenkins wrote:

  Dublin Core:
  Phill currently has issue with them regarding duplicate or confusing
  standard for document language which conflicts or at lease duplicates with
  html lang= attribute.

Received on Tuesday, 15 January 2002 17:00:09 UTC