Associating DSSSL style sheets with documents

[Edward C. Zimmermann]

> 
> Within the next week or so, I expect to announce the availability of a
> Web server that can respond to certain kinds of URLs with an XML data
> stream (eventually, a variety of XML data streams).  For our own
> design purposes, and also for the purposes of experimenters working
> with combined XML/DSSSL applications, I would like to see this group
> come up with an unofficial method or methods by which to associate a
> DSSSL style sheet with a particular chunk of XML.  Such methods would
> be far in advance of the sgml-wg specification effort and subject to
> later revision, but given the influence of experimental
> implementations, I think that it's appropriate to put a little bit of
> thought into the design.
> 
> One possible method suggested by James Clark (thank you, James) is to
> adopt the convention used by Jade in the absence of the -d option:
> replace the extension of the document entity's URL or file name with
> .dsl and fetch that.  Thus, if a browser fetches
> 
>    http://docs.sun.com/foo/bar.html
> 
> then it should also look for
> 
>    http://docs.sun.com/foo/bar.dsl
Since no public DTDs must have the DTD, viz a URL to DTD.. and
from the name/path/URL to DTD then one can use the extension
.dsl or whatever for the DSSSL.
The problem with using .dsl as the map from the URL .extension is
that if one has a class of documents built around a DTD and that
has a DSSSL "style sheet" then one will either need to have a
front-end server to manage this whole bit (why) or fill the place
with symbollic links.. The problem with both are that proxy caches
will get filled with redundant bits... Why not use the DTD URL as
base? Either this or I don't understand what your aims are, or
I'm totally lost:-)
The other alternative, of course, would be to extend HTTP to
return a request for the association of a URL to the .dsl from
a file.. That is probably the better and more flexible way but
it won't work with popular off-the-shelf browser and thus is
ill-suited to experiments......
> 
> and apply it to bar.html if found.
> 
> This is appealingly straightforward, but I wonder how well it
> accommodates multiple stylesheets and stylesheets that use other
> notations (CSS, for example).  Of course, we could deal with the
> second concern by saying that DSSSL is the default stylesheet language
> for XML experimentation and that we will figure out some way to
> accommodate other stylesheet languages later.
> 
> James lists some other possibilities:
> 
> | - a processing instruction somewhere in the prolog
> | 
> | - a catalog entry that says unconditionally to use some DSSSL style
> |   sheet
> | 
> | - a catalog entry that associates a DSSSL style sheet with the public
> |   identifier of a DTD
> | 
> | - make the document serve also as a style sheet by making it conform
> |   to the DSSSL architecture (this will work with Jade too)
> 
> Any thoughts on this?  I am, of course, particularly interested in
> hearing from those of you who are actively building DSSSL
> applications.
> 
> Jon
> 
> 
> xml-dev: A list for W3C XML Developers
> Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/
> To unsubscribe, send to majordomo at ic.ac.uk the following message;
> unsubscribe xml-dev
> List coordinator, Henry Rzepa (rzepa at ic.ac.uk)
> 
> 


-- 
______________________
<A HREF="whois://rs.internic.net/ecz">Edward C. Zimmermann</A>
<A HREF="http://www.bsn.com/">Basis Systeme netzwerk/Munich</A>

xml-dev: A list for W3C XML Developers
Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/
To unsubscribe, send to majordomo at ic.ac.uk the following message;
unsubscribe xml-dev
List coordinator, Henry Rzepa (rzepa at ic.ac.uk)