On Wed, Jul 23, 2003 at 06:57:06AM -0700, Rudi Cilibrasi wrote: > I'm glad to see there's interest here.
Yes indeed, this stuff definitely looks interesting.
Very shortly Matthew will be creating the API for the class that will
encapsulate the event recording and estimation data for a single node.
Rudi - perhaps you could outline your idea of what this should look like
for a ML algorithm.
Basicially, I think it should look something like:
public class NodeEstimator {
public void registerDataFound(BigInteger key, long startTime, long duration) { ... }
public void registerDataNotFound(BigInteger key, long startTime, long duration) {
... }
public void registerRouteNotFound(long startTime, int type, long duration) { ... }
public long estimateRetrievalTime(BigInteger key, long now) { ... }
}
Am I missing anything?
Anyway, I want to get the current algorithm implemented to this API,
this will give us a baseline against which we can evaluate refinements
or replacements for it.
Ian.
--
Ian Clarke [EMAIL PROTECTED]
Coordinator, The Freenet Project http://freenetproject.org/
Founder, Locutus http://locut.us/
Personal Homepage http://locut.us/ian/
pgp00000.pgp
Description: PGP signature
