[I hope this gets auto-forwarded to the moderator]
The discussion about caches is interesting. I get increased speed at both
the hgmp cache and the hensa cache.
I was interested in whether there is an argument of using the hgmp cache as the
people who connect are all likely to be molecular biologists and so it is more
likely to be cach'ing the documents of interest to us.
ie - is it sensible to have a "local" cache in the sense of local information
content. The fact that HENSA cache has xxx pages and only needs to get pages
at 1% of the time etc is somewhat irrelevant to us as - to be honest - we are
likely to be the 1% rather than www.playboy.com
I guess this question is all about cache size - does HENSA have to chuck out pages
due to space limitations or not? If this is a yes then maybe a "information local"
cache makes sense
does anyone else have any ideas on this?
ewan birney
