- From: Karl Dubost <karl+w3c@la-grange.net>
- Date: Fri, 20 Feb 2009 19:35:05 -0500
- To: Rob Sayre <rsayre@mozilla.com>
- Cc: "John Foliot - WATS.ca" <foliot@wats.ca>, "'Geoffrey Sneddon'" <foolistbar@googlemail.com>, "'HTML WG'" <public-html@w3.org>, "'W3C WAI-XTECH'" <wai-xtech@w3.org>
Le 20 févr. 2009 à 19:10, Rob Sayre a écrit : > I do agree that validators can encourage people to do things, or > alert them to problems they weren't previously aware of. I'm curious > how a validator would deal with the requirements in Ian's draft for > the alt attribute. I also know of a very successful validator that > manages to inform users of problems not quite covered by spec > requirements. Here's an example: > > <http://feedvalidator.org/docs/warning/NotInline.html> Yes very good validator. Olivier has been recently working on fuzzy matching for the W3C Markup Validator. http://www.w3.org/blog/systeam/2009/02/16/validator_fuzzy_match > So, while I agree that specs beget validators, I disagree that > validators need RFC2119 imperatives to do their job in this case. I didn't say that. :) >> "Not working well at this point in the message…" I think I >> understand your goal, but "reality" and "real world" should be >> banned from all discussions we have about technologies. These >> expressions are a smoke screen to the actual issues > > There are certainly W3C Working Groups that appear to have banned > those terms. I don't think it is a good idea. Let me put back the part you cut ;) Le 20 févr. 2009 à 18:30, Karl Dubost a écrit : > and often a flag to promote the idea of one's group on the line of > "My reality is real, not yours." :) The issue is that talking about reality add a layer of semantic indirection in the discussion. > <http://code.google.com/webstats/> > Is a good survey. was a good survey. You should look at the awesome work that Brian Wilson has recently published. http://dev.opera.com/articles/view/mama/ Btw There is a call for suggestions for MAMA 2.0 http://www.iheni.com/ask-mama-what-the-web-is-made-of/ > Looks like there are a lot of alt attributes out there. I wonder how > many are empty, and how many are valid according to the requirements > in Hixie's draft (difficult to measure!) Yes very difficult to measure. It needs often human checking. Plus the evaluation of a useful alt will again depends on the context.
Received on Saturday, 21 February 2009 00:36:32 UTC