Subject: Re: [xsl] Copy results|
From: Steven Ericsson-Zenith <steven@xxxxxxxxxxxxx>
Date: Tue, 8 Jul 2008 04:38:30 -0700
Obviously I am using multiple stylesheet modules (though not 50 at this point), and indeed can generate in part or whole. The documents themselves contain type information to determine the stylesheet choice. Is Ant used to regenerate the whole (all the standards you mention) or per each logical document?
In that particular case some steps in the pipeline are the same for all
documents (for example, extraction of all the defined terms and section ids
into a single master index document), while some steps apply only to one
document (for example, creation of the XQuery and XPath documents from a
common master); other steps are similar for different documents but with
different overlay stylesheets (for example, the XSLT spec has an overlay to
define the formatting of element syntax templates). For most of the
documents a diff document is generated by a stylesheet that compares with
the baseline version, but for some the editors maintain diffs by hand.
Encoding the workflow within the documents themselves is a valid approach,
especially if different documents can be at different stages in their
lifecycle. But in that case you've designed your own miniature pipeline
language, and you still need an interpreter for it... You can write that in
XSLT (I've done so in the past), but you again hit problems as soon as you
want to insert tasks into the pipeline that aren't natural XSLT tasks.
Michael Kay http://www.saxonica.com/
-- Dr. Steven Ericsson-Zenith Institute for Advanced Science & Engineering http://iase.info http://senses.info