On Tue, 2003-08-05 at 19:38, Ian Clarke wrote:
> On Tue, Aug 05, 2003 at 07:30:40PM -0400, Scott Young wrote:
> > With the testability of NGR, we could have a programming contest to find
> > the best algorithm for NGR, sort of like Google's programming contests. 
> > A fund could be set up as an award for coming up with the finally chosen
> > algorithm.  This type of competition would hopefully increase Freenet's
> > PR and bring in more donations, as well as speed up development of the
> > routing algorithm in the process.
> > Comments?
> 
> An interesting idea - although I am not sure that people really need to 
> have a financial incentive for this, there are no shortage of people 
> willing to do this kind of research just because it is fun and 
> interesting.

True, but a financial incentive would be more of an encouragement. 
Think of all the college students that need some money to pay off their
loans.


> Of course, the first thing is to get the NGR infrastructure in-place 
> with the default implementation.

If we want to set up a fund, we could start it now.  Publicity on
Slashdot would likely bring donations, developers, and contestants.

One more important aspect of a competition is that there be a way of
arbitrating which algorithm works best.  It could be a simple
measurement, or it could be a decision made by one of the developers.  A
servlet with some statistical analysis of the algorithm's performance
would be useful.

I also think it would be good to separate the routing algorithm code
from freenet.jar and have a separate Jar for the routing code so that a
specific build doesn't have to be compiled once for each routing
algorithm.  Testers could take build x and plug in routing algorithm y
without having to find a specific build xy.



_______________________________________________
devl mailing list
[EMAIL PROTECTED]
http://hawk.freenetproject.org:8080/cgi-bin/mailman/listinfo/devl

Reply via email to