: I explicitly asked on java-user once, "Hey, how does/can this thing : work....blah blah", but got no responses. As far as I know, Chuck is : the only ParallelReader users. :)
my understanding is that ParallelReader only makes sense if you truely are building both indexes in sync -- keeping the adds/deletes/optimizes in perfect unison -- and the advantage comes only from the ability to have searchers which span the "full" index using the Parallel reader *or* do do searches on a single reader. so it might make sense for having stored fields that you only use some of hte time, or TermEnums that you only use some of the time -- and in the apps where you aren't using them, they aren't loaded into ram. -Hoss
