[
Lists Home |
Date Index |
Thread Index
]
On Saturday 04 December 2004 18:39, Andrzej Jan Taramina
wrote:
> Do you happen to know if anyone has tried to build an
> XSLT engine that does incremental transformations on
> incoming SAX events, without requiring the building of a
> tree? That kind of approach, where the transform is
> appropriate, would be much more efficient in memory
> useage and would allow transforms of virtually unlimited
> size documents I should think. Something to
> investigate...
I don't think this will be of much practical use to you but
just to add more to the picture. As far as I am aware
Sarvega is the only XSLT vendor who has been pushing on in
this area. A few months ago we achieved client defined
transforms on documents over the 4GB boundary on standard
PC hardware. Our theoretical limit is 1024GB although I
don't think we will be trying that for a while yet.
Once you get beyond memory limitations the problem is of
course CPU time which for many stylesheets is some function
of input size. To scale to these levels you need to have a
stylesheet that at worst exhibits linear CPU time wrt
document size. This is just another way of saying that
performance is highly dependent on what your stylesheet
does.
The main challenge we have ahead for very large handling is
to extend the optimisations we perform to allow execution
over a greater range of stylesheets. This is also a goal
for smaller documents as well so it should just happen over
time.
Kev.
|
