You may have read "a few" as meaning "1 or 2", but that's not what I
wrote. I was suggesting the heuristic "if you've got a 64Mb machine
don't try to process more than about 6Mb of source data."
I just wanted to point out that this is a really common source of
problems for people, and the heuristic depends on the scenario. For
example, if you have a 500MB of RAM, and are transforming only a 1MB (on
disk) file inside a servlet, you are still going to be in big trouble
once you get a certain number of simultaneous connections against the
web server. The main thing that people need to understand is that
resource consumption for the typical in-memory XML cache is much greater
than the size on disk of the XML. That's where most people get into
trouble -- they prototype with small XML and few connections, and things
work fine. But they never realize that their memory bloat is a factor
of 5x or 10x until the run out of memory.
XSL-List info and archive: http://www.mulberrytech.com/xsl/xsl-list