Re: Media types vs. conformance/performance?

On Sun, 1 Oct 2000, Bjoern Hoehrmann wrote:
>
> For example
>
>   <link rel="stylesheet" media="aural" href="foo.css" />
>
> Where foo.css is like
>
>   @import url("tv.css") tv;

I believe that the HTML and CSS specs are intended to make tv.css in the
above scenario never be used. As I understand it a media-dependant @import
is only useful in a stylesheet that applies to all media. In the case
above, foo.css only applies to aural media, and tv.css only applies to
aural media AND tv media -- but since they are mutually exclusive, that
means that tv.css is never used.


> This is an invalid statement. Authors must define media="all" if rules
> apply to all media types. The default value of the media attribute in
> (X)HTML and so XML is "screen".

This is an acknowledged error in the spec. The default should be 'all' not
'screen'.


> Now let's say I'm using a screen media user-agent and tell it to print
> the document. What media type is the user-agent of now?

Print (and only print).

The media types (with the exception of 'all') are mutually exclusive.


> This leads me as an author to say I always have to declare media="all"
> if the style sheet is not very special for a certain media.

I strongly agree; IMHO almost all stylesheets should apply to all media.


It has been proposed that a system such as CC/PP be used to extend media
types in a way such as described in the HTML4 spec as a potential
improvement. I don't know what the status of this proposal is now.

-- 
Ian Hickson                                     )\     _. - ._.)       fL
Netscape, Standards Compliance QA              /. `- '  (  `--'
+1 650 937 6593                                `- , ) -  > ) \
irc.mozilla.org:Hixie _________________________  (.' \) (.' -' __________

Received on Friday, 5 January 2001 11:14:51 UTC