- From: Robert Burns <rob@robburns.com>
- Date: Wed, 15 Aug 2007 18:05:21 -0500
- To: Alfonso Martínez de Lizarrondo <amla70@gmail.com>
- Cc: "Philip Taylor" <philip@zaynar.demon.co.uk>, public-html <public-html@w3.org>
HI Alfonso, On Aug 15, 2007, at 5:46 PM, Alfonso Martínez de Lizarrondo wrote: > > 2007/8/16, Robert Burns <rob@robburns.com>: >>> >>> So in order to understand if they have tried to use something >>> correctly you have to manually test each of those pages and find out >>> if their use is correct or not, not just assume that it could be a >>> valid hypothetical use of a feature. >> >> I understand there are all sorts of crazy things on the web. But the >> point is Phillip's research show this is not implemented in any UA. >> Yet were finding authors trying to use it. > > That assumption is wrong. > > You don't know if anyone of them really tried to use it, or it's just > something that somehow did get inserted automatically or by someone > that was testing things. > > How do you explain the existence of pages with a "name" attribute set > on table cells: > <http://canvex.lazyilluminati.com/survey/2007-07-17/analyse.cgi/ > tagattr/td/name>? > > You need to look at the pages to know if they really tried or the page > is just a lot of tag soup. The assumption may be wrong or it may be right. The point is, it just doesn't matter. Let's assume that all of these authors were just typing randomly and happened to type <input type='image' usemap='*'>. Then the number of authors trying to use this HTML 4.01 feature that has not been implemented is 0%. OK, then that's what I would expect it to be, as I've been saying all along. However, if we discuss this as a WG, I would think it would be one of those hypothesese that's just not worth testing. Finding out that no one has implemented it, is some interesting research. Its also research that's typically easier and less costly (in terms of time and effort) to conduct. Once we know that, I can't imagine why we'd want to learn how many authors have used the feature even though its unimplemented or simply by accident. That we find use of an unimplemented feature I find remarkable. Its not the type of research I think is useful, but as trivia its remarkable.. I really don't think the authors simply accidentally typed <input type='image' usemap='*'>. Whether it is used precisely as the HTML 4.01 recommendation suggests or some other way, I still would think it is related to the HTML 4.01 recommendation. Even authors who read the recommendation rely heavily on implementation indicators to verify that they're using the feature correctly (with this feature they have no way to do that). So to sum up, in my view: 1) conducting research to learn that this has never been implemented is a useful; 2) after learning that a feature has never been implemented, conducting statistical research to learn whether the feature is used by authors even though it has never been implemented is not at all useful; 3) having conducted the research in (2) anyway, we learn nothing by checking to see whether authors used the feature "correctly" based on some hypothetical UA that might have implemented the feature in the way we imagine it should have been implemented. Once conduct research (1), I think a different research path is called for. Take care, Rob
Received on Wednesday, 15 August 2007 23:05:38 UTC