I'm interested in this mainly as an academic exercise, because I see using reflection as the only dynamic behaviour that Java offers and it allows for some 'tricks' (for want of a better phrase) to write less code. I'm also interested in the sort of reception that this kind of code gets - most people who I've worked with (Java programmers) hate it, but some people (Perl programmers) think it's a nice solution to a problem.
Reflection is not bad in itself, but be aware that it's gotten a lost faster in newer JDK, so your perf test on JDK 1.5, which shows no slow down, might do terribly on JDK 1.2 for example. I suspect Java jocks avoid reflection more because of the ugly syntax than any other reasons (at least with using a recent JDK with fast reflection). --DD PS: No I don't have access to the trees any more. I'm -1 on making DirScanR slower for sure ;-) Regarding this change, I'm -0 until tested on a 10/15 level deep hierarchy, but with thousands of files, with some dirs having a large number of files. I think these were the "crucial" caracteristics of the trees I had. --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
