xml-dev - Re: xslv updated

[Rick JELLIFFE <ricko@geotempo.com>]

Sebastian Rahtz wrote:
 
> Following DSSSL, they decouple the stylesheet from the
> formatter. Which makes it impossible to implement
> 
>  1) a design which says "captions are centered unless they are longer
>     than one line, in which case they are ragged right"
> 
>  2) a cross-reference system which says "figure 1 on the opposite
>     page" or "figure1 on p.56", depending on where it floats to
> 
> two things which are easy enough to do in LaTeX (to take a typical
> benchmark)
 
> But if the XSL cannot talk to the formatter, how can it base other
> style decisions dependent on what the page floating came up with?
 
In the current draft, one could treat both the cases above as
"multi-case"s, where your formatter provides the feedback to select the
correct sequence of areas, I think.  (Of course, there has to be some
check in place to make sure that an area toggled on here is not toggled
off elsewhere. But that is not difficult.)

For the first example, it would be possible also for XSL to specify it
declaratively (if they so extended XSL) to allow a
justification-if-more-than-one-line property.  Or you could specify this
property yourself, as an extension on top of standard XSL (which your
own formatter would understand).

For the second example, I suppose for batch processing one could also
use the formatter's indexer and perl/XSLT to generate a document
containing the page number reference for every citation in the document,
along the lines
  <ref at="4" to="9" >p. 9</ref>
  <ref at="5" to="6" >p. 6</ref>
  <ref at="6" to="7" >p. 7</ref>
then perform an XSL tranformation so to
  <ref at="4" to="9" >p. 9</ref>
  <ref at="5" to="6" >the following page</ref>
  <ref at="6" to="7" >the opposite page</ref>
and read the data in as external references. (Of course, this would take
a few runs to stabilize, but that would not be a big change for TeX or
batch people.)

Have you sent in comments on this to the XSL FO group? They may be able
to provide properties (these uses you mention seem to gracefully
degrade, so having extra properties does not seem to add complexity to
people who don't implement them: the basic properties would hold.)

Rick Jelliffe

***************************************************************************
This is xml-dev, the mailing list for XML developers.
To unsubscribe, mailto:majordomo@xml.org&BODY=unsubscribe%20xml-dev
List archives are available at http://xml.org/archives/xml-dev/
***************************************************************************