SAX, expat and JNI

[David Megginson]

Joerg Brunsmann writes:

 > assuming that interpreted Java byte code is still slower than
 > compiled C code, it is desirable to use James' expat XML Parser in
 > conjunction with SAX to gain maximum performance while parsing XML
 > documents.

Surprisingly, this is an assumption that will not always hold true.
On a P166 NT box using Microsoft's jview, for example, AElfred can
chew through about 1MB of XML each second.  I have not tested XP on an
NT box, but since it is about 10-20% faster than AElfred on my Linux
notebook, it could be even faster on the NT box as well.

In other words, unless you're dealing with very large XML documents
(say, 10MB or more), you'll probably lose more time loading the DLL
and converting data in the stubs than you'll gain in having a slightly
faster parser.

My experience in the past has been that at least 90% of your
application's time will be spent processing the information delivered
by the parser; in fact, you would probably do better to leave the XML
parser in Java and implement the event handlers in C.

That said, SAX is non-language-specific precisely so that people can
do the sort of thing that you propose, and I'll be very interested in
following your progress.


All the best,


David

-- 
David Megginson                 ak117 at freenet.carleton.ca
Microstar Software Ltd.         dmeggins at microstar.com
      http://home.sprynet.com/sprynet/dmeggins/

xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev at ic.ac.uk
Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/
To (un)subscribe, mailto:majordomo at ic.ac.uk the following message;
(un)subscribe xml-dev
To subscribe to the digests, mailto:majordomo at ic.ac.uk the following message;
subscribe xml-dev-digest
List coordinator, Henry Rzepa (mailto:rzepa at ic.ac.uk)