recursion in XML parser

[roddey at us.ibm.com]

> > I think that nobody would argue that Java has a lot of virtues that
> > certainly speed of not one of them. To take your numbers David, If
> > that part of the application is 10 times faster than any Java
> > parser and that the app itself is 10 times faster also. The overall
> > throughput is therefore 10 times faster.
>
>That's not necessarily the case -- C/C++ have some advantages for fast
>I/O that Java doesn't share, but if your other code is not I/O-bound,
>and if it doesn't require small, tight processing loops, the speed
>difference for the non-parsing code might be much less significant
>(depending on how efficient your VM and OS are at memory-management).
>

And startup time is an issue as well. If you must parse lots of little
files, the overhead of Java's frameworks and the problems with
initialization of data and whatnot can cost you getting out of the gate.
We've noticed significantly better 'get it up, get it out, get it down'
performance on C++, which can be a concern in the e-bidness area.

Another issue we've noticed is that object creation (regardless of the
downstream issues of cleaning them up) has a pretty darn high cost in Java,
making it difficult to do high performance code that is as OO as one might
wish sometimes. If performance is the Holy Grail, and it seems to be the
primary measurement made (probably because its easy to measure, not because
its what's important), then there can be a pull to write less maintainable
code to gain performance. A language with less overhead for use of objects
(at both ends of the the lifecycle) can sometimes do better in that regard.



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/ and on CD-ROM/ISBN 981-02-3594-1
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)