XOM API for Processing XML with Java

[Wolfgang Hoschek <whoschek AT lbl.gov>]

> > 3) Verifier.checkPCDATA
> > FYI, there are still Unicode calculations (multiplications, etc) in 
> > decodeSurrogatePair that are entirely unnecessary. Actually the 
> > entire method is unnecessary. My fastest version is as follows 
> > (retaining semantics unchanged):
>
> I see what you're getting at here, but I find it really hard to 
> believe decodeSurrogatePair has any practical impact. This is only 
> going to come up if the text contains characters from beyond the basic 
> multilingual plane. Does Lawrence Berkeley Labs really do that much 
> processing on Tengwar and Cirth? :-)

Who knows, a few years from now we might be communicating with aliens 
from hidden corners of the universe, using the highest planes. Assuming 
that they use a concept of symbols. Of course, it might not happen, 
perhaps they'll see our use of character encodings as a compatibility 
threat to some kind of holy fetish, in which case they might deem our 
"preventive destruction" a more practical avenue, leaving little room 
for negotiating a standard process. Just kidding :-)

> I suspect the speed-ups you see here are based on using charAt instead 
> of toCharArray, not eliminating the extra operations in 
> decodeSurrogatePair. This is the sort of thing that is very VM 
> specific, and perhaps even document specific. The reason I'm using 
> toCharArray here is that in my tests on both Linux and Mac OS X this 
> was faster. I can easily believe this could have turned around again 
> in Java 1.5 or some other variation.

If there's a difference it stems from the use of a more or less 
optimizing compiler. I'm using JDK 1.5 server VM. Apps runnning a 
client VM in production presumably don't need or care about 
optimizations anyway.

Wolfgang.