RE: Intent to Conform (was Re: Version information)

L. David Baron wrote:
On Thursday 2007-04-12 11:01 -0700, Chris Wilson wrote:
> >[on <!DOCTYPE html5> not triggering standards mode]
>> Umm, it does in IE.  And I'd expect it to in any other browser,
>> too - it's an unknown doctype.
>It's an incorrect DOCTYPE declaration (since the name given doesn't
>match the root element for the document format).

Hey, you've read an SGML spec.  :)  Of course, we all thought it was a good idea not to be bound by SGML, right?  The point made about XML by someone is a good one, though.

>Would you expect "<!DOCTPE html>" (missing "Y") to trigger standards mode?

Nope, because we look for <!DOCTYPE.  We just don't compare to the root element name.  A good point, though.

>What about "<!html>"?  What about "DOCTYPE" without any markup around it?

Ditto above - we look for <!DOCTYPE...>, so no and no.

>Perhaps that was a mistake, but the basic idea of Mozilla's quirks mode selection code (which I wrote) is:
> 1. Parse the doctype.  If we can't (e.g., if it's not present or
>    malformed), use quirks mode.
>... This triggers quirks mode because it fails in step (1).

Apparently the determination of malformed is not quite the same across Mozilla, Opera, Safari and IE.


Received on Friday, 13 April 2007 17:12:42 UTC