On Wed, 20 Jan 2010, Maciej Stachowiak wrote: > > I haven't seen any follow-up discussion. I'm interested in hearing what > the rest of the Working Group thinks. Does anyone strongly agree with > Matt that the sentence he objects to should be removed? Does anyone > strongly feel that the sentence should be retained? Does anyone have > alternate wording to suggest that might be acceptable to everyone? I think that its's important that we at least acknowledge the possibility that user agents use image analysis techniques, and certainly that we explicitly allow the use of such techniques. Even a straightforward OCR of many images with no alternative text would dramatically improve the accessibility of many pages. I don't really understand what the problem with the sentence in the spec now is. If anything, I'd say the text is a bit weak - given recent advances in image analysis techniques (e.g. as seen in Google Goggles), we might in fact want to consider changing the "may" to a "should". I do not at all buy the assertion that Google is somehow uniquely positioned to perform image analysis -- there are open source OCR packages available that can target decade-old hardware that perform quite well. It has been argued that this implementor-targetted sentence might in some way counter the pages and pages of detailed author-targetted text that the spec has encouraging, nay, requiring, that authors include alt="" text. I do not believe this to be the case, but would be happy to include further text in the spec admonishing thinking such thoughts. HTH, -- Ian Hickson U+1047E )\._.,--....,'``. fL http://ln.hixie.ch/ U+263A /, _.. \ _\ ;`._ ,. Things that are impossible just take longer. `._.-(,_..'--(,_..'`-.;.'Received on Thursday, 21 January 2010 06:39:29 UTC
This archive was generated by hypermail 2.4.0 : Saturday, 9 October 2021 18:45:08 UTC