|
|
Posted by John Dunlop on 06/30/05 19:39
Michael Winter wrote:
> On 30/06/2005 08:44, John Dunlop wrote:
>
> > I can see one workaround [for unrecognised types:] Write a custom DTD
> > in which the <type> attribute of SCRIPT is optional.
[By the way, I should've mentioned, I don't approve of my
suggestion at all. Let it be known, I'm in no way
recommending it!]
> I don't really see that working, though. Browsers don't actually read
> the DTD, so they're unlikely to care whether you want the type attribute
> to be treated optionally.
Well, that was really for the purposes of validation, to get
round SCRIPT's obligatory <type> attribute.
> Moreover, and as I said to Ian, an omitted type attribute is different
> from a specified, but unrecognised, MIME type value.
Yes. In my suggestion the <type> attribute is left out. The
only place the soon-to-be MIME type features is in the HTTP
Content-Type header.
> The former could permit a reasonable guess (some ECMAScript
> derivative), but the latter must cause a script to be ignored.
So, in your words, as it falls into the former category, my
suggestion 'could permit a reasonable guess'?
PS, where does the spec say, other than in informative notes -
and I'm *assuming* the note in sec. 6.7 is informative - that
the <type> attribute takes a value of a MIME type as defined
in RFCs2045-49? Are normative references normative if they
occur in informative notes?
--
Jock
Navigation:
[Reply to this message]
|