It's not really a problem for me, but I would like to prove (or least
test) that it'd be a good idea to use an XSLT solution in a GUI
environment like OpenOffice for this sort of thing.
I'm not sure how you did your timing, but when doing a one-off transform
with saxon often the bulk of the elapsed time is spent starting up the
JVM.
Mike's posts on this list often advise people to call saxon via its API
calls from a compiled java program so that you don't start up the JVM
for each document (and can often avoid recompiling the stylesheet for
each document as well) although actually I don't tend to do that
even if running the same stylesheet over a thousand documents I tend to
do it as as a loop in a command shell (bash)
for i in */*.xml
do
saxon ....
done
The fact that it takes a bit longer isn't an issue here (it takes long
enough to have lunch anyway, and it's done when you get back) and it
means that you can easily customise your loops from the command shell
even on systems that only have a run time system and not a full java
development environment (and by people who don't speak java)
I don't know if OOo has a JVM permanently running that you can connect
to, or whether the input and output filters need to start up the java
virtual machine each time?
david
________________________________________________________________________
This e-mail has been scanned for all viruses by Star. The
service is powered by MessageLabs. For more information on a proactive
anti-virus service working around the clock, around the globe, visit:
http://www.star.net.uk
________________________________________________________________________