docbook-apps

As I went digging through the XSLT 2 stylesheets again this morning,
after a reasonable absence, I became more and more convinced that
they're a mess architecturally.

It all works in its own clever way, but the fact that overriding the
root template involves deducing that you actually have to override

<xsl:template match="*" mode="m:root">

strikes me as setting the bar awfully high for your average
customizer. Doubly so when you consider that overriding "/" causes the
whole edifice to fall over.

I'm tempted, with an eye towards an XProc-aware future, to rip it all
apart and rewrite it as an explicit sequence of discrete
transformations instead of a mode-driven, internal sequence of
transformations.

Thoughts?

Be seeing you,
norm

--
Norman Walsh <ndw@nwalsh.com> | A man can believe a considerable
http://www.oasis-open.org/docbook/ | deal of rubbish, and yet go about
Chair, DocBook Technical Committee | his daily work in a rational and
| cheerful manner.--Norman Douglas

Norman Walsh wrote:

> I'm tempted, with an eye towards an XProc-aware future, to rip it all
> apart and rewrite it as an explicit sequence of discrete
> transformations instead of a mode-driven, internal sequence of
> transformations.

I'd hesitate to state it outright, but the chances are with
the advanced features of both XSLT 2.0 and Relax NG schema,
there's likely to be lots of room for improvement Norm?

Are there any particular aspects to 2.0 you think you'd
like to use, or is it just the effects of n years wear and tear?


More thoughts. If the group is interested in letting more people
look to modding the stylesheets, the architecture could stand
being documented somewhat.

I'm thinking about the sort of thing that Jeni T does, presenting
the stylesheets as a navigable suite. Unfortunately her software is
Windows only.
http://www.jenitennison.com/xslt/utilities/ xsltdoc



regards

--
Dave Pawson
XSLT XSL-FO FAQ.
http://www.dpawson.co.uk

Norman Walsh wrote:

> I'm tempted, with an eye towards an XProc-aware future, to rip it all
> apart and rewrite it as an explicit sequence of discrete
> transformations instead of a mode-driven, internal sequence of
> transformations.

I can understand to your motivations for moving to XProc on the other
hand relying on XProc will mean that we are raising bar for user -- they
have to install not only XSLT2 processor but also XProc processor. It
also mean that it will not be possible to use stylesheets directly in
tools like oXygen, XML Spy and StylusStudio which supports XSLT2, but
not XProc.

But I can imagine that it should be possible to generate XSLT stylesheet
which will chain discrete transformations in the same way as XProc does.
If such stylesheet will be distributed as a part of stylesheets together
with XProc pipeline I think that it will satisfy all interested parties.

Jirka

--
------------------------------------------------------------------
Jirka Kosek e-mail: jirka@kosek.cz http://xmlguru.cz
------------------------------------------------------------------
Professional XML consulting and training services
DocBook customization, custom XSLT/XSL-FO document processing
------------------------------------------------------------------
OASIS DocBook TC member, W3C Invited Expert, ISO JTC1/SC34 member
------------------------------------------------------------------