"Michael Kay" <mhk(_at_)mhk(_dot_)me(_dot_)uk> wrote in message
news:000001c2effb$be3c7c60$6401a8c0(_at_)pcukmka(_dot_)(_dot_)(_dot_)
Mike, common subsequence sharing really makes a big
difference. For example list sharing is one of the major
assumptions in estimating the time complexity of the algorithms in
Thanks for the advice, it's certainly one of the things I will put on my
list of future opportunities. I can't immediately see how to combine it
with lazy evaluation, but I'll stew on it for a while.
It seems to me that list sharing does not interfere with lazy evaluation at
all. It comes as part of an operation only when the decision has been
already taken to really perform that operation.
On the other side, when a decision is taken that an operation must be
performed, this is done because the result is necessary and regardless (or
even not knowing) how the operation will be performed.
I've recently extended the optimizations I do in two particular areas -
compile-time moving of sub-expressions out of loops, and run-time lazy
evaluation, and between them they have made a big difference to
execution time for some complex stylesheets. The bad news is that most
stylesheets are not complex enough to benefit from them!
Therefore, there should be a way (e.g. command line switches) for the user
to turn on such optimizations when they are really needed.
=====
Cheers,
Dimitre Novatchev.
http://fxsl.sourceforge.net/ -- the home of FXSL
XSL-List info and archive: http://www.mulberrytech.com/xsl/xsl-list