Re: CSS/XHTML Validators weirdness

On 16.07.01 at 01:07, Bjoern Hoehrmann <derhoermi@gmx.net> wrote:

>>  <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN"
>>  "DTD/xhtml1-strict.dtd">
>
>If there is no http://xtrmntr.org/salo/SILC/DTD/xhtml1-strict.dtd the
>CSS Validator is right here, if there is, the Validator will issue the
>same message, too :-)

You need to set "OVERRIDE YES" in your SGML Declaration for that FPI;
you are prefering the System Identifier over the FPI.


>>  <!DOCTYPE HTML PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN"

s/HTML/html/;

The error message is misgiving.


>There is no statement, that an unresolveable SYSTEM identifier renders a
>document not-well-formed or invalid.

It doesn't. The System Identifier, by definition, is system specific. It's
assumed to be meaningfull within the context of the application, but not
required to be universally resolvable.


>The W3C HTML Validator however doesn't catch this, because he uses a
>catalog system to find local DTDs rather then fetching them from the Net.

It prefers to use a catalog and an Entity Manager to resolve the External
Subset for certain Public Identifiers, but will use the System Identifier
for those it doesn't know.

Received on Sunday, 15 July 2001 20:16:00 UTC