Re: Invalid pages...

On Wed, 14 Feb 2007, Bachu wrote:

> why is that in w3.org has 7% of pages ie 14 pages
> didnt pass the w3 validation?

Please clarify. There are surely more than 200 pages at w3.org, so 7% is 
surely more than 14. Do you mean you checked 200 pages? How? I did not 
find any "validate multiple pages" option in the W3 validator; is there 
one?

Well, there _is_ such an option in the WDG validator, and using it that 
way,
http://www.htmlhelp.com/cgi-bin/validate.cgi?url=http%3A%2F%2Fwww.w3.org&spider=yes&hidevalid=yes
(validating "entire site", which is effectively just 100 first pages found 
by the validator, and suppressing warnings and OK messages)
indeed reveals that there is something rot...invalid at w3.org. There were 
5 pages with errors, i.e. 5 % invalid pages among the tested. That's 
surely better than web average, but...

(Some of the warnings are actually relevant too, e.g. warning about the 
use of XHTML syntax in HTML.)

> is this some error in validator or in the pages ?

In the cases I tested, the errors reported are genuine errors on w3.org 
pages, not bugs in the validator, and probably the W3C validator would 
report them as well:
- missing alt attributes in <img> elements
- illegal ("non-SGML" or "non-XML") characters
- missing document type declaration
- bogus document type declaration <!doctype html>
- omitted end tags in XHTML
- malformed tags where attribute contents appears where attribute
   name is expected
- uppercase in attribute names in XHTML

(and some spurious error messages caused by fundamental errors at the 
start of document, triggering a mode of parsing other than the intended 
one).

This isn't about the validator, really. Rather, about W3C quality control, 
though specifically about not using W3C's own tools. But it affects the 
credibility of propagating the use of the validator.

-- 
Jukka "Yucca" Korpela, http://www.cs.tut.fi/~jkorpela/

Received on Thursday, 15 February 2007 13:12:53 UTC