enhancement request: limited version history

Hi-

I've got 2700 java files (tons of html too) with 5 yrs history,
recently imported from CVS.

I like to scan incoming changes, and check over why the changes were made.
Options for making this fast:
1) con SVK to act as a local write-through SVN, caching all rev.
history. At least the fetches are local and not over the net.
2) have subclipse cache all fetched revisions. After all, it's a
read-only data set.
3) limit fetched history to, say, 3 months or 20 revisions, whichever is less.

1) needs the command-line interface to SVN to be completely finished.

2) is most attractive, but I can see there would be memory size problems.
my workflow would be:
  a) synchronize the whole code base
  b) fetch revision history on whole code base, caching.
  c) look at rev history on each file as needed, fetching cached versions.

3) would work pretty well, also, assuming revision fetch walks
backward from HEAD through the SVN db.

Thanks for your interest.

--
Bring back Nixon!