Dori a écrit:
On 6/11/05, Tim Starling t.starling@physics.unimelb.edu.au wrote:
Ævar Arnfjörð Bjarmason wrote:
Tim reconfigured it to have one vote per bug and a very large pool of votes ( given our # of bugs ) per user, each user now has one vote per bug.
I like it.
Now you can actually tell which bugs are more wanted than other bugs. Some people were voting 1, some 100, some 400, some 700. When you add them all together you get a number which has not much relation to anything. There were 3 developers present on IRC, all were in favour. Bugzilla sent an email to everyone who had voted telling them that their votes had been declared invalid, bless it.
I like this as well. It makes it a lot easier to figure out at a glance how many people want a certain feature/bug.
Well, voting methods always have benefits or drawbacks.
Now, you can measure how many people really support a feature. But no more.
Previously, you could mostly measure how much a feature was important to those who voted for it. But you could also go see details and see how many people really wanted it.
In a sense, results are easier to read now for you as they have only one interpretation (number of supporters), while before they had two interpretations (number of supporters AND degree of support). They are easier, but far less informative. Perhaps a middle solution might have been to allow between 1 to 3 points for each feature.
This said, as several mentionned here, it does not really matter actually, as developers are volunteers and work on features they feel like working on.
Recently, a journalist reminded me about wikimoney and asked me why I think it failed. Well, most of us needs at the same time to work on something that is fun for them and something useful for others. Possibly first driven force is more important than second ;-)
Ant