On 11/22/08, Jürgen Ragaller <[EMAIL PROTECTED]> wrote: > Am 19.11.2008 um 23:49 schrieb [EMAIL PROTECTED]: > > Voting: I prefer a system allowing every person to vote on every > > entry. Only allowing positive votes for five entries will not give an > > accurate reflection of preferences. What if we only receive 4 > > entries? What if we receive 100 entries? What if 6 entries are very > > similar? Voting for every entry will discover which entries are most > > acceptable. Once the number is reduced to a manageable (2 or 3, > > definitely less than 10) quantity, a typical vote-for-favorite system > > can determine the winner. > > > > Vote scores: The Hate-Love example phrases were demonstrating the > > concept. Better words should be used for the contest: > > Unacceptable - Dislike - Good - Fantastic > > Mmh - for my feeling that is still a bit problematic. I prefer a system > with "points" or "stars" (maybe feathers?). > > > > > > Mailing List vs. Application: This contest is an opportunity to > > demonstrate Lenya. An application to collect votes both eliminates > > the work of counting the votes and demonstrates how Lenya is useful in > > the real world. (Of course, the negative is someone needs to build > > the application and add it to our website.) > > > > While the idea is somewhat appealing I share Andreas' thought, that > ml-voting will be sufficient. > > > Vote confirmation: Each voter must provide an email address. The > > email address is sent a memo requiring a response to validate the > > vote. The response may be a link confirming the vote, a reply, or the > > traditional password for login. > > > > > > Vote Strength: We could score votes based on the voter. This would > > be difficult without an application. Thoughts are: > > Unacceptable - Dislike - Good - Fantastic > > - Lenya Committer: -9, -5, 5, 9 > > - Apache.org email: -7, -4, 4, 7 > > - Lenya ML Posters: -5, -3, 3, 5 > > - Everybody else: -2, -1, 1, 2 > > > > If we'd use stars, that could transform into something like: > > 10 stars for committers > 8 stars for apache.org emailees > 6 stars for ml posters > 4 stars for everybody else > > In addition to that I find it necessary that as a last resort, vetoing > (suggestion: two -1 ones from committers) is possible. > And I'd like to add a note the the voting process can be cancelled by the > committers in case of less than 4 (or 3) entries or when there is obvious > misuse in the voting process (from more than one person). > > Thanks again and have a nice weekend > Jürgen
I changed the Wiki to remove public negative votes. Still prefer voting for every entry. Example using every entry of 15 non-vetoed entries: Fantastic counts as 3 votes for up to 1/5th of the entries = 3 entries = 9 votes Acceptable counts 1 vote, assume half of remainder = 6 entries = 6 votes No vote - assume remainder = 6 entries.= 0 votes = 15 votes Giving 15 votes to be distributed at the voter's discretion could have one person assigning 15 votes to a single entry while another splits votes between two similar entries. If we allow distribution of a certain number of votes, we should set a maximum number per entry. How about number of votes equals number of entries with a maximum of one-fifth for a single entry? People could vote for their favorite five entries or spread votes to all acceptable entries. This is still not as optimal as voting separately on each entry. Also changed Wiki to reflect ASF rules that project changes require an official vote. Rule check: Are binding votes from all Committers or only PMC members? solprovider
