On 6 Oct 2009, at 1:17am, P Kishor wrote: > Basically, I want to > imply more confidence in the calculated rank for an "idea" if the > number of people who voted for it is higher... so, more the number of > votes for an "idea," more the confidence in its rank.
Not quite. You could have lots of votes for an idea, but everyone could give it a different number of stars. Then you would have a rank (the average of the votes) but your confidence in it would be very low. Another idea might get only five votes. But if everyone who voted for it gave it the same score, you'd have a lot of confidence in that score. Another workable solution. What you're really talking about there is standard deviation. A small standard deviation means everyone gave the idea roughly the same score. A large one would mean that the votes were all over the place. On 6 Oct 2009, at 1:17am, P Kishor wrote: > Yes, a particular vote can be changed, but of course, only by its > voter. So, if I gave 10 stars to an idea today, I can come back > tomorrow and change it to 8 stars. In that case, you need to have an UPDATE trigger as well as an INSERT trigger. And in that case it becomes very important that you create a record in your votes table for the original ten stars the creator gives the idea. Simon. _______________________________________________ sqlite-users mailing list [email protected] http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
