RE: ACTION-982: style sheets

Surely a prohibition on transforming style sheets is only relevant if
the content that they refer to has not itself been transformed? Given
caching considerations etc. how can one assess whether this is the case
or not?

Jo

> -----Original Message-----
> From: public-bpwg-request@w3.org [mailto:public-bpwg-request@w3.org]
On
> Behalf Of Eduardo Casais
> Sent: 19 June 2009 20:07
> To: public-bpwg@w3.org
> Subject: ACTION-982: style sheets
> 
> 
> ACTION-982
> 
> Propose some specific text ref ISSUE-298 (style sheets).
> 
> The fundamental idea is to avoid irrelevant transformations of
> resources that are
> unambiguously identified as mobile-optimized. Style sheets are a case
> in point. I
> already mentioned that there are proxy deployments that strip out
style
> elements
> (inline or external) marked as handheld without any particular
> justification.
> 
> I propose to add the following text in 4.2.9 after the very first
> bullet point:
> 
> -------
> o  the content is CSS and within the scope of a @media declaration
> explicitly
>    associated with the handheld media type;
> o  the content is an included resource explicitly associated with the
> handheld
>    media type;
> -------
> 
> The first bullet point takes care of two situations:
> 1.  inline and internal style rules, bracketed by @media handheld, ...
> { ...rules...}
> 2.  CSS rules contained in an external style sheet and themselves
> bracketed as in 1.
> In both cases, the CSS in question is self-descriptively marked as
> mobile-specific.
> 
> The second bullet point takes care of two further situations:
> 2.  external style sheets referred to from an HTML/XHTML document:
>     <link rel="stylesheet" media="handheld,..." href="..."
> type="text/css" />
> 3.  external style sheets referred to from another sheet:
>     @import url("...") handheld, ...;
> In both cases, the referred resource is explicitly identified via a
> link as mobile
> specific.
> 
> Given the haphazard state of CSS implementations in mobile devices, it
> is impossible
> to rely on the first technique (@media) only; as a matter of fact,
> several mobile
> browsers handle @media incorrectly or not at all. Describing the
> properties of a
> resource via the link to it is the safest way and it is standards
> compliant. That
> this mechanism was not standardized to take into account
transformation
> proxies
> may be regrettable, but it is no justification not to try making sure
> that mobile
> optimized sites are respected. Analysing content and then deciding to
> retrieve
> resources depending on the media type is already considered in the
> guidelines in
> section 4.2.7 -- and many operations of modern transcoders imply some
> form of session
> anyway. Finally, this of course does not prevent operations such as
> compressing style
> sheets, correcting syntax errors and the like.
> 
> 
> E.Casais
> 
> 
> 
>

Received on Monday, 22 June 2009 14:59:22 UTC