- From: David Woolley <david@djwhome.demon.co.uk>
- Date: Sat, 2 Jul 2005 10:04:56 +0100 (BST)
- To: www-style@w3.org
> (if money is involved). If the computer can do it, let it. Let's find > a way for computers to do the accessibility and portability work for > us. Expecting the content author so far hasn't been too successful. Both of these are difficult AI problems as they involve things like character recognition of deliberately difficult text, deep semantic understanding of both text and images, combined with a knowledge of the laws and codes of practice on what can be legally put in real print (in order to provide summaries of images that are appropriate in context), etc. Because web pages are usually, in part, advertisements, they are very subtle means of communication and it is very difficult to mechanically extract the intended psychological impact (it can also unacceptable to the author to explicitly encode this). Actually, with alternative text, it would be better, from a semantics point of view to make the text the primary content when the image is used for text replacement, in which case the AI problem becomes that of working out the house style and generating appropriate text as image. Even if they can be mechanically derived, they still have to be part of the document standard because they have to go over the wire in order to work on the old browsers, or on the slow machines that may be all that is available on a mobile phone or in a poor country. (Some people argue that accessibility only applies to those that are considered legally disabled, I consider it about making content available to as many people as possible.)
Received on Saturday, 2 July 2005 21:46:24 UTC