On Wed, 12 Nov 2008, Andrei Alexandrescu wrote: > Frits van Bommel wrote: > > Andrei Alexandrescu wrote: > > > Brad Roberts wrote: > > > > Ok, enabled for the D product. Each registered user has 10 votes and > > > > can vote only once per bug. > > > > > > What is the motivation behind limiting the total number of votes a user > > > may have? I think that may blunt the statistics a bit. > > > > Presumably it's to force you to only vote for the ones you consider most > > important. > > This way you may not be able to vote for every issue you ever ran into and > > instead have to pick the ones that bug you most (no pun intended). In > > theory, this could lead to more meaningful results, especially as time > > passes. > > Hmmm... If a user has unlimited votes they can't manipulate the system. Say > they vote 1 for every single bug. That simply means an additional bias is in > the system, which does not influence the ranking. > > So the limit of 10 is really an arbitrary cap on the number of bugs a given > person can consider important. > > As time passes, this only gets worse. Say I "consumed" my 10 votes. Then I > find a showstopper for my code. Well I need to decide which of the other bugs > I don't "care" for anymore. Why do I need to make that decision? > > Don't introduce arbitrary aperture windows in the system. Let the statistics > do its work. > > Andrei
Do you have more than 10 you want to vote for? It's not a cap on how many are important. It's a cap on how many you can suggest are the ones that time should most urgently be spent on. Anyway, like I said, I'll change it if there's sufficient interest and so far there clearly hasn't been. :) Later, Brad