On Thu, 13 Jun 2002, Phil Harris wrote:

Dennis Allison wrote:

...

The minidom won't quite do what I need. This particular project is a web-hosted XML editor, really a development front-end for another system. The ParsedXML product holds and manages the corpus of the material being edited...

Now that I have a semi-working alpha-release it may be time to rethink the system and redesign for performance and simplicity. ParsedXM drags along a lot of mechanism that I do not need. In fact, my experience as been that XML and the DOM is not a great internal representation--better to parse the XML into a more maleable data structure, do what needs to be done, and reconstruct.

B> Reading this mail I just thought of this:

http://www.reportlab.com/xml/pyrxp.html

It's advertised as "pyRXP - the fastest XML parser?".

Thanks for the pointer. I'll take a look at it -- I use the PyXML minidom for much of my non-Zope processing. I do not know whether the performance problem I am having is parsing related or associated with all the cruft that comes with the ParsedXML implementation. When I get a moment to breath I'll instrument and see... My guess is that it's not the parser, but guesses about performance problems are notoriously flakey.