Thanks, I didn't take that into consideration. Although I still couldn't get anywhere close to 20ms, but it's getting down well below a sec.
Thanks again for the advice. Andy Clark wrote: > Alan Sung wrote: > > All I did was to run the DOMCount demo. If I'm not mistaken, the > > DOMCount takes the time before the parsing, then the time after > > parsing(and traversing the DOM tree, but that doesn't take too > > long with a small to medium size file) and subtract them. Then it > > would be just the run time excluding the JVM start up time. > > Yes, but the first parse in the same VM *is* including the time > required to dynamically load all of the classes needed by the > parser. I always call the DOMCount program twice in the same > invocation so that the first parse brings in the classes and I > can more readily believe the second time. For example: > > java dom.DOMCount -v document.xml document.xml > > > Is there something wrong with my setup then? All I did was java > > -classpath . dom.DOMCount "XMLfile" -v (for validation) within the > > Xerces directory. > > Options must appear before the files on the command line that > you want them to take affect for. > > -- > Andy Clark * IBM, JTC - Silicon Valley * [EMAIL PROTECTED]
