W3C home > Mailing lists > Public > w3c-wai-ig@w3.org > July to September 2002

RE: Support for LONGDESC

From: Charles McCathieNevile <charles@w3.org>
Date: Tue, 6 Aug 2002 11:19:05 -0400 (EDT)
To: Jon Hanna <jon@spin.ie>
cc: <w3c-wai-ig@w3.org>
Message-ID: <Pine.LNX.4.30.0208061111560.25450-100000@tux.w3.org>

A couple of thoughts:

If you describe an image that you created for your site, it makes sense to
link it to places it is included on your site.

If the image is used like a Web thing (i.e. it has a single URI, rather than
being in multiple places) then you can of course search for that URI as an
image included in other pages. I think google does this, otherwise you could
do it with the Webbot, although it is always risky to program your own
web-crawl...

Alternatively there is work on image annotation that allows for multiple
descriptions of the same image. This is the kind of thing RDF systems are
good for. An example would be looking at Annotea
http://www.w3.org/2001/Annotea - an annotation system that lets you describe
something, including an image, where you can have multiple descriptions
created for the same document, or image (or even, if you use SVG, part of an
image). Some more information on image annotation is available in the form of
a preliminary report draft (including links to various tools and projects) on
a workshop about image annotation -
http://www.w3.org/2001/sw/Europe/reports/dev_workshop_report_1/

cheers

Chaals

On Tue, 6 Aug 2002, Jon Hanna wrote:

>
>> > Also, as the longdesc file is handled as a standard HTML file using HTTP
>> > by the browser, it is liable to be indexed by indexing spiders, so
>> > potentially could be found as a result of a search query, without
>> > appropriate context, navigational bars, etc.  (I appreciate that it can
>> > be useful to find such information - I'm trying to think through the
>> > implications).
>>
>> Again could be handled a couple of different ways, via robots.txt,
>> .htaccess, <meta name="Robots" content="noindex,nofollow" />, etc.  All
>> three if you're really paranoid! <grin>
>
>There goes the baby as well as the bathwater!
>
>Presumably if a search engine has matched a description of an image to a
>search query then that image in some way satisfies the search. Indeed such
>discovery for a search engine would likely be more reliable than is
>available from most sites (which is why the image search is probably the
>least dependable part of those search engine sites which offer one). The
>text of a longdesc file can be considered metadata about the image after
>all.
>
>Rather than try to discourage such indexing, take advantage of it:
>
>1. Provide navigation that makes the longdesc useful as a "first page" in
>the site.
>2. Provide machine-readable links to every document that you know uses, or
>is very likely, to use the image, so the search engine can associate the
>image with the original document:
>
><head profile="http://purl.org/DC/elements/1.0/">
>	...
>	<link rel="schema.DC" href="http://purl.org/DC/elements/1.0/" />
>	<meta name="DC.Relation.IsRequiredBy" content="blah.html" />
>	<link rel="DC.Relation.IsRequiredBy" href="blah.html" />
>	...
></head>
>(or perhaps DC.Relation.IsPartOf or DC.Relation.IsReferencedBy as
>appropriate), so there is at least a better chance of the search engine
>associating the image it now has a description of with its primary context.
>

-- 
Charles McCathieNevile    http://www.w3.org/People/Charles  phone: +61 409 134 136
W3C Web Accessibility Initiative     http://www.w3.org/WAI  fax: +33 4 92 38 78 22
Location: 21 Mitchell street FOOTSCRAY Vic 3011, Australia
(or W3C INRIA, Route des Lucioles, BP 93, 06902 Sophia Antipolis Cedex, France)
Received on Tuesday, 6 August 2002 11:19:06 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 19 July 2011 18:14:05 GMT