Natterings about history files
davidsen at tmr.com
Tue Mar 13 17:56:33 UTC 2001
In article <7xf6hBQmw-B at khms.westfalen.de>,
Kai Henningsen <kaih at khms.westfalen.de> wrote:
| davidsen at tmr.com (bill davidsen) wrote on 02.03.01 in <97n8cf$g9h$1 at gatekeeper.tmr.com>:
| > I believe that there is a flavor of perfect
| > hash which gives you either a hit with value or a miss,
| That's pretty much the definition of a perfect hash. But perfect hashes
| are useless for fast-changing data; they're extremely compute-intensive to
| set up, in exchange for being really cheap to look up stuff in.
That's why I thought it would be useful if we ever go to a multi-file
history, with no new entries being added to any copy but recent. That
would make the lookups very fast for the old files, useful if you have
to check several files.
One more thing we could do to scale INN at some point.
bill davidsen <davidsen at tmr.com>
CTO, TMR Associates, Inc
Doing interesting things with little computers since 1979.
More information about the inn-workers