Quoting Toad <[EMAIL PROTECTED]>:I would think that's the effect the original poster was getting at, the preferred node (lowest pDNF) being continually tried, knocking it (pDF) down until the same as the rest. I wonder if it's the same for others.
We have a mechanism called "probability of legitimate DNF", which should compensate for most of this noise.
Most (if not all) the time, "probability of legitimate DNF" is the same as "probability of DNF" on my node.
The current estimate is using pDNF - pLegitDNF, which for J means DNFs aren't a factor.
Neither is pLegitDNF*tLegitDNF counted.
tLegitDNF maybe/maybe-not the same as tDNF, but in the current code it must contribute.
----
pLegitDNF is and should be the same across all nodes. So take my next point as an imprecise (but hopefully useful) thought.
----
Perhaps pLegitDNF (and tLegitDNF) could be calculated per node using;
-when all queries for the key fail each (DNFing node) will get their pLegitDNF increased.
-when a latter node finds data previous nodes get there DNF increased.
Maintaining pLegitDNF+pDNF+pDF == 1
Pick an updated estimate function.
Unfortunately it would be a bit warped with the wizardry (decreasing HTL) used currently.
_______________________________________________ Devl mailing list [EMAIL PROTECTED] http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/devl
