On 25 Apr 2007, at 21:08:09, Sean Fraser wrote:

>> The first ranked most popular Alexa site, Yahoo!, adresses your
>> points above.
>>
>> W3C HTML Validation = 33 errors [http://validator.w3.org/check?
>> uri=http%3A%2F%2Fwww.yahoo.com%2F]

>However this is misleading, in that Yahoo! uses server-side sniffing
>to serve different content to different user agents.

The error numbers from URI Submission are misleading when compared with Direct Input/Form Submission error numbers when using W3C HTML Services. The HTML5 Conformance Checker found a fatal error in Line 21, Column 136. [http://hsivonen.iki.fi/validator/html5/?doc=http%3A%2F%2Fwww.yahoo.com] Regardless, Yahoo!'s home page failed. That isn't  HTML4-happiness.

>I get the same
>33 errors for HTML validation if I simply give the validator the URL
>of Yahoo!'s home page, but viewing the source used by the validator
>shows that it is completely different to the source sent to my copy
>of Firefox, Safari or IE.

Interesting.

>Interestingly, using the FF
>Web Developer extension's "Validate local HTML" option gives 272 HTML
>validation errors.

That's nearly identical to direct input of Yahoo!'s home page source code as seen with PC/FF; W3C HTML Validator returned 273 errors.

>Perhaps this could be addressed by allowing users of the validator to
>specify a particular UA header to be used when retrieving a page for
>validation.

>[1] <http://developer.yahoo.com/yui/articles/gbs/>

I doubt that most users would specify a particular setting. Last year, when CSS Validation discussions occured, most users of the CSS Validator did know know that they could change the default setting. Consequently, the CSS Validator Team made the default setting 2.1.

The survey I propose would - simply - have Pass/Fail criteria; HTML5 would include error numbers and issues.


--
Sean Fraser
http://www.elementary-group-standards.com