I dont think you should be so quick to over-ride that maintainer's evaluation. It would seem likely that they have looked into it a bit deeper than the average user. (Maybe weight the maintainers rating)
Also may move to marks out of ten rather than ill-defined garbage...gold. I would doubt anything under wine merits gold and garbage mean many things to many ppl.
Using less emotive number ratings allows more objective ratings and finer granularity (0-10 not 0-3) .
Equally one digit takes up a lot less space than "garbage" and would allow display of both maintainer and user ratings which is clearly more informative.
HTH .
On Fri, 28 Oct 2005 19:08:00 +0200, Chris Morgan chmorgan@gmail.com wrote:
This sounds like a plan. If Tony is ok with it we'll schedule to get the maintainer rating removed and replaced by an average of the maintainer rating for the latest wine rev probably middle or late next week. I wanted to get the first pass of the unit testing code in place before we make any more changes so we can reduce the risk of breaking the production site.
Chris
On 10/28/05, Jonathan Ernst Jonathan@ernstfamily.ch wrote:
Le vendredi 28 octobre 2005 à 12:36 -0400, Chris Morgan a écrit : [...]
Right now the rating you see is the one decided by the maintainer of the application and based on their judgement. Now that we have the testing results there is some question of how to reflect these results to users. Certainly if an application has a handful of gold ratings and no maintainer rating it should probably be listed as gold. Do we remove the maintainer rating all together? Add yet another field? Ideas? Personally I'm inclined to remove the maintainer rating and have the application rating be based on some computation based on testing results but there is a question of whether those are reliable enough and whether we want the maintainer to be able to override for some reason.
I think the maintainer rating should be removed and the last rating should be used for version rating. If the version is maintained the maintainer can submit test results on a regular basis and thus influence the overall version rating. -- Jonathan Ernst Jonathan@ErnstFamily.ch
-----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.1 (GNU/Linux)
iD8DBQBDYlZ25uNwURKeRvURAsCOAKCA+b3E5P6zeFv3BfZyiViNYmQ2BACeJ+an K1mXBmF/jW8A0iLLz6aLCvM= =/p/l -----END PGP SIGNATURE-----