On Fri, Jun 20, 2003 at 09:34:50PM -0400, Ed Tomlinson wrote: > On June 20, 2003 08:03 pm, Ian Clarke wrote: > > I am not sure I see what this has to do with NGrouting exactly - am > > I missing something? > > It adds/hooks together some of the needed infrastucture. The next thing > for me to do is look at feeding the response time estimator class some info > and seeing what sort of results we get from it. > > There is also quite a bit more stats on what is happening while routing. For > instance we now track (but do not use) the normalized time for sucessful > queries both globally and per node. This and the message success rates > is a good judge on how well we are routing.
Ok, are you comfortable with the algorithm I proposed for dealing with connection failures and such-like? The whole idea of the NGrouting algorithm is to avoid alchemy, I am somewhat concerned that you just checked in a whole bunch of alchemy - I would rather start with something basic that conformed to the NGrouting "way of doing things" and then add stuff from there - rigorously justifying any alchemy if it is absolutely required. Perhaps I am misunderstanding what you just checked in, 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
