- From: Arjun Ray <aray@q2.net>
- Date: Tue, 28 Sep 1999 03:20:40 -0400 (EDT)
- To: www-html@w3.org
On Mon, 27 Sep 1999, Liam Quinn wrote:
> Bert Bos wrote:
> > 4) Section 2 says that doctype can be used to distinguish the
> > versions.
To be meaningful, this has to be a statement about the public and/or
system identifiers, not the doctype declaration. Even so, the public text
class of the FPI is DTD; hence the entity addressed is just the bunch of
declarations that are logically and syntactically an integral part of the
document instance.
However, this is not versioning in the true sense, because that also
involves - even if only in the sense of an acknowledgment - the normative
prose of the relevant spec. An FPI for that would have public text class
NOTATION, not DTD, and anyway, the doctype declaration is not where you
put such a reference. (In fact, there is no determinate way to do this at
all in ISO 8879.)
> > It should probably also say something about the absence of a
> > doctype, which is allowed at least in HTML 2.0.
>
> This is often asserted, but I don't believe that it's true. From
> <http://www.w3.org/MarkUp/html-spec/html-spec_3.html#SEC3.3>:
>
> +To identify information as an HTML document conforming to this
> +specification, each document must start with one of the following
> +document type declarations.
>
> Note the use of "must".
I believe Bert is refering to this, from the same section:
NOTE - If the body of a `text/html' message entity does not begin
with a document type declaration, an HTML user agent should infer
the above document type declaration.
> > Maybe it should say that the absence of a doctype means version 4.01
> > (or the "latest version" in case there are any after 4.01). Or maybe
> > that the absence of a doctype means that all bets are off...
Actually, given the state of 99.99% of documents shoveled onto the wire as
'text/html', I'd rather they didn't include doctype declarations at all.
Arjun
Received on Tuesday, 28 September 1999 02:42:16 UTC