- From: Ian Hickson <ian@hixie.ch>
- Date: Tue, 14 Aug 2007 21:19:11 +0000 (UTC)
- To: "Philip Taylor (Webmaster)" <P.Taylor@Rhul.Ac.Uk>
- Cc: public-html <public-html@w3.org>
On Tue, 14 Aug 2007, Philip Taylor (Webmaster) wrote: > > So are you arguing that the design of HTML 5 should be based, in part, > on a putative "need" to support sites that have used a documented > feature of HTML 4.01, yet which rely on that feature not working as > documented in some fraction of their visitor's browsers? Are we really > going to re-define the syntax and semantics of HTML based on the > contra-normative use of it by a small number of seemingly perverse > sites? Oh yes, definitely. As you say, it's only a part of the many things we should base our research on, but in some cases (such as <input usemap>) it becomes a critical part. It's not necessarily a small number, either; usemap on <input> is used on hundreds of thousands of pages, yet in almost all cases (87% in the study I did) the attribute points to nowhere or to a <map> that has no useful <area> elements. (And, as noted before, the remaining dozen thousand cases or so are limited to a few hundred sites, and most of those, by inspection, work better in browsers that don't support <input usemap>.) -- Ian Hickson U+1047E )\._.,--....,'``. fL http://ln.hixie.ch/ U+263A /, _.. \ _\ ;`._ ,. Things that are impossible just take longer. `._.-(,_..'--(,_..'`-.;.'
Received on Tuesday, 14 August 2007 21:19:30 UTC