On June 29, 2003 07:24 pm, Ian Clarke wrote: > > P_best_dnf()*dnf_time(node,key)*htl+(P_dnf(node)-P_best_dnf())* > > global_success_estimate(key) > > You aren't taking into account the time required for the first node to > DNF before re-requesting. > > This solves that problem: > > P_best_dnf()*dnf_time(node,key)*htl+(P_dnf(node)-P_best_dnf())* > (dnf_time(node,key)*htl + global_success_estimate(key))
If I factor correctly this works out to: P_dnf(node)*dnf_time(node,key)*htl+ (P_dnf(node)-P_best_dnf())*global_success_estimate(key) Still like it? > > I will also look at putting a switch that allows ng to be turned on > > and off. > > Indeed - I would like to get this in unstable ASAP switched off by > default (or at least somewhere that it can be more widely tested). Problem is that, no mater what, it will change routing as the code is all based on the stuff I previously posted to devl. My personal preference would be to sync experimental to unstable and update it with the code. Ed _______________________________________________ devl mailing list [EMAIL PROTECTED] http://hawk.freenetproject.org:8080/cgi-bin/mailman/listinfo/devl
