- From: Terje Bless <link@tss.no>
- Date: Fri, 3 Mar 2000 00:12:12 +0100
- To: W3C Validator <www-validator@w3.org>
On 02.03.00 at 16:59, Christian Smith <csmith@barebones.com> wrote:
>On Thursday, March 2, 2000 at 14:10, hugo@w3.org (Hugo Haas) wrote:
>
>>I don't think that the validator checks the validity of a URI.
>
>Actually, since it's a DTD based validator I don't believe there is -any-
>way for it to check the validity of a URI with respect to encoding.
Sure there is. The "href" attribute to the "A" elementis defined to be
"%URI" which is in turn defined to be "CDATA": "a sequence of characters
from the document character set" that "may include character entities".
That is one check that can, and should, be made. If it passed the HTML you
passed it, it's a bug and should be fixed.
(
In addition, there is no reason why we can't add additional checks to
make sure that the document follows the Recommendation and not just the
DTD. The Validator isn't supposed to just Pretty Print the output of
SP, it's supposed to check the document for actual Validity. How
feasible that is in practice is another matter and each extra check
needs to be examined carefully before implemented. Maybe the check
should be foisted on Weblint or be made an optional feature not
directly related to the validation pass.
)
However, "g:foo" is perfectly valid. Just because it isn't a recognized
scheme, doesn't make the syntax invalid. This is a XML/SGML validator; not
a URI validator.
Received on Thursday, 2 March 2000 18:34:23 UTC