Fw: About the style processing instruction

["Oren Ben-Kiki" <oren@xxxxxxxxxxxxx>]

Didier PH Martin <martind@xxxxxxxxxxxxx> wrote:



><MyComment>
>If this is the case, don't we have to define a clear algorithm which
>stylesheet is the "best" match? This would be trivial if the user specifies
>a single back end. However if the user to specifies several alternatives
>(say with a priority for each) then it becomes much less simple.
>
><Reply>
>Right to the point! Yes Oren, I think you are right, it would be a good
>thing to have the algorithm explicit. I'll think about it. Do you have any
>suggestion for this?


Anything I come up with in an hour of thought is sure to be wrong :-) The
problem is that you have to tradeoff between the priority the user has given
each back end, and the quality of the match of this back end to any given
stylesheet. This isn't in general solvable so "everyone" would be happy;
that's why a simple rule should be specified so at least we'll be able to
rely on its quirks :-) Of course the user could always request manual
selection of the back-end, just in case...

><Reply>
>You mean to use the MIME type instead of the single type. Yes, I think this
>would be a good idea, IANA already list all registered MIME types. So, If I
>understand you well, the PI would be like:
><xml-stylesheet href="MyScript.dsl" type="text/xsl"
>media="screen,application/tex"?> instead of
><xml-stylesheet href="MyScript.dsl" type="text/xsl" media="screen,text"?>
>Right?


Yes, except that it isn't that simple. This assumes that there's a 1-1
matching between a <Media type, MIME type> to a back-end. I'm not too happy
with this assumption, it is used in browsers, mail programs, OS-es and so
on.

Share & Enjoy,

    Oren Ben-Kiki


 XSL-List info and archive:  http://www.mulberrytech.com/xsl/xsl-list