[jdom-interest] My use of JDOM

[Philipp Gröschler]

Iván Martel wrote:
> As a poor-man XMLDB. To integrate various Elements in a caching system 
> and perform rapid XPath searchs, and cached XSLT transformation of 
> those results to obtain great (<1ms) performance. 

I am very interested in the part where the caching comes in. Could you 
please explain a little more with what techniques and in which environment 
(especially CPU power and the XSLT transformer you're using) you gain that 
kind of performance?

Currently I'm dealing with the problem that my code using JDOM is very poor 
on speed (counting in seconds, even with very small documents), typical 
operations are like merging two document trees (with recursively traversing 
one until finding the right Element and then attaching the second tree) or 
stuffing XML into a transformer. As I want to introduce caching, I'm using 
JDOM objects all over (JDOMSource to feed the transformer).

So do you have a few tips how to gain more speed? Does caching bring in so 
much and do you have to deal with thread concurrency? (meaning, how do you 
use a bunch of elements in a document tree without cloning or locking them 
out from other threads?). Do I have to use a different transformer 
(currently Xerces-J 2.6.0) or simply am I to buy a new machine? :-)

Thank you so much,

Philipp