Re: several messages

On Fri, 29 Jun 2001, Chris Lilley wrote:
> Ian Hickson wrote:
>> On Wed, 27 Jun 2001, Chris Lilley wrote:
>>> Making it optional means that editing tools have no way to figure out if
>>> a stylesheet is CSS or XSL other than by attemptig to resolve the URL.
>>> This is not a good idea.
>> How else do you propose linking to a document without knowing its MIME
>> type in advance?
> Are you arguing for or against, here .....

I'm arguing for the status-quo of the "type" pseudo-attribute being
compatible with HTML4, just like the rest of the XML stylesheet PI spec
and in fact as intended by the author according to the first paragraph of
section 1 ('This processing instruction follows the behaviour of the HTML
4.0 <LINK REL="stylesheet">').

> yes exactly, how is an authoring tool to know unless that markup says
> so directly.

By checking. (In fact, an authoring tool rarely needs to know.)

This has never been a problem with HTML as far as I am aware; why would
it suddenly become a problem with XML?

If I have a CGI script which sniffs for the UA string and returns XSL for
IE5, CSS for Mozilla and JSSS for Nav4, there is no way I could link to it
using a specific MIME type. Ergo, the "type"  pseudo-attribute, which
being a useful optimisation for many UAs, cannot be a required
pseudo-attribute in the real world.

> I notice that XML Spy 4.0 for example assumes that all xml-stylesheet
> PIs are for XSL, and removes any existing CSS ones if it adds an XSL
> one, with the message 'replacing existing XSL style sheet"

In which case it is _definitely_ buggy...

Ian Hickson                                            )\     _. - ._.)   fL
Invited Expert, CSS Working Group                     /. `- '  (  `--'
The views expressed in this message are strictly      `- , ) -  > ) \
personal and not those of Netscape or Mozilla. ________ (.' \) (.' -' ______

Received on Friday, 29 June 2001 08:56:19 UTC