All,
So it seems to be almost every other day now when somebody mentions how broken the appdb rating system is. This "broken-ness" has had several high level discussions on Wine-Devel. There have also been many, many proposed solutions to this problem.
I do not wish to discuss these problems or these solutions in this thread. It is time something was done about this. By done I mean agreed upon (probably the harder part) then implemented in code and published to the winehq server(s).
I'de like to arrange an IRC discussion for all those interested in finalizing an attack plan, and then assigning the responsibility of doing it to somebody who is capable and has the time (perhaps myself).
Lets (arbitrarily) set a time for the chat at:
When: 11PM UTC (6PM EST) 2/6/09 (Friday) Where: #winehackers
If this arbitrary time seems to be poorly chosen for a large number of people then please propose an alternative well in advance so all are aware of it.
--Zach
Zachary Goldberg <zgold <at> bluesata.com> writes:
All,
So it seems to be almost every other day now when somebody mentions how broken the appdb rating system is. This "broken-ness" has had several high level discussions on Wine-Devel. There have also been many, many proposed solutions to this problem.
I do not wish to discuss these problems or these solutions in this thread. It is time something was done about this. By done I mean agreed upon (probably the harder part) then implemented in code and published to the winehq server(s).
I'de like to arrange an IRC discussion for all those interested in finalizing an attack plan, and then assigning the responsibility of doing it to somebody who is capable and has the time (perhaps myself).
Lets (arbitrarily) set a time for the chat at:
When: 11PM UTC (6PM EST) 2/6/09 (Friday) Where: #winehackers
I think this will be a very difficult if not impossible task to accomplish. If we talk about test results, it's the app maintainer that is responsible for the correctness of the ratings and test results. What would you want to do if you think that the ratings for application X are overestimated? Remove the maintainer from the application entry in appdb? Then you'd have to check yourself for every app, if ratings are correct.
BTW, also the opposite occurs very frequently: applications that are rated too low (garbage) whereas they are probably bronze or higher. When users send in garbage test results, i'm always tempted to try out the app myself if it has free download. In many cases it appears the application needs additional packages like mdac, vbrun or anything. Also many garbage results come from installing in a busted wineprefix ( just do a winetricks dcom98, and you'll get garbage results for apps in many cases), and they run just fine in a new ~/.wine
So if someone shouts: "Hey, the app is rated gold, but for me it crashes" it's not necessarily the "gold-rater" who's wrong.
My conclusion really is, it's very difficult to get the ratings correct, but always worth a try of course. What's more important is I think , that the appdb is rich source of info on how to get application X or Y running.
Regards
So my conclusion is it will be very difficult to get
On Tue, Feb 3, 2009 at 4:12 PM, Louis Lenders xerox_xerox2000@yahoo.co.uk wrote:
Zachary Goldberg <zgold <at> bluesata.com> writes: I think this will be a very difficult if not impossible task to accomplish. If we talk about test results, it's the app maintainer that is responsible for the correctness of the ratings and test results. What would you want to do if you think that the ratings for application X are overestimated? Remove the maintainer from the application entry in appdb? Then you'd have to check yourself for every app, if ratings are correct.
These problems could be mitigated by making it easier to submit test results. Even something as simple as a 5-star "how well does it work for you" system that any registered user can submit with a single click. Yes, the information we get will be less specific, but we will get a *LOT* more of it. I would rather know that 1740 people have gotten a game to work and 400 haven't (and no other information than that) than to see just just five detailed test results saying gold/gold/silver/silver/garbage.
2009/2/3 Sparr sparr0@gmail.com:
On Tue, Feb 3, 2009 at 4:12 PM, Louis Lenders xerox_xerox2000@yahoo.co.uk wrote:
Zachary Goldberg <zgold <at> bluesata.com> writes: I think this will be a very difficult if not impossible task to accomplish. If we talk about test results, it's the app maintainer that is responsible for the correctness of the ratings and test results. What would you want to do if you think that the ratings for application X are overestimated? Remove the maintainer from the application entry in appdb? Then you'd have to check yourself for every app, if ratings are correct.
These problems could be mitigated by making it easier to submit test results. Even something as simple as a 5-star "how well does it work for you" system that any registered user can submit with a single click. Yes, the information we get will be less specific, but we will get a *LOT* more of it. I would rather know that 1740 people have gotten a game to work and 400 haven't (and no other information than that) than to see just just five detailed test results saying gold/gold/silver/silver/garbage.
As I mentioned previously this thread's purpose is not to decide how this is to be done. Thats what the IRC meeting is for.
We can go on and on about how its best to be done over wine-devel but unless the people who have a stake in the matter and the ability to do it sit down and make some hard decisions the talk won't result in action.
Zachary Goldberg wrote:
All,
So it seems to be almost every other day now when somebody mentions how broken the appdb rating system is. This "broken-ness" has had several high level discussions on Wine-Devel. There have also been many, many proposed solutions to this problem.
I do not wish to discuss these problems or these solutions in this thread. It is time something was done about this. By done I mean agreed upon (probably the harder part) then implemented in code and published to the winehq server(s).
I'de like to arrange an IRC discussion for all those interested in finalizing an attack plan, and then assigning the responsibility of doing it to somebody who is capable and has the time (perhaps myself).
Lets (arbitrarily) set a time for the chat at:
When: 11PM UTC (6PM EST) 2/6/09 (Friday) Where: #winehackers
If this arbitrary time seems to be poorly chosen for a large number of people then please propose an alternative well in advance so all are aware of it.
--Zach
Zach, I completely agree with this strong-armed measure. Let's get it done. :)
Thanks, Scott Ritchie
On Tue, Feb 3, 2009 at 9:44 AM, Zachary Goldberg zgold@bluesata.com wrote:
All,
So it seems to be almost every other day now when somebody mentions how broken the appdb rating system is. This "broken-ness" has had several high level discussions on Wine-Devel. There have also been many, many proposed solutions to this problem.
I do not wish to discuss these problems or these solutions in this thread. It is time something was done about this. By done I mean agreed upon (probably the harder part) then implemented in code and published to the winehq server(s).
I'de like to arrange an IRC discussion for all those interested in finalizing an attack plan, and then assigning the responsibility of doing it to somebody who is capable and has the time (perhaps myself).
Lets (arbitrarily) set a time for the chat at:
When: 11PM UTC (6PM EST) 2/6/09 (Friday) Where: #winehackers
If this arbitrary time seems to be poorly chosen for a large number of people then please propose an alternative well in advance so all are aware of it.
--Zach
I have some thoughts on the subject. I'll try to be there
Branan
On Tue, Feb 3, 2009 at 11:44 AM, Zachary Goldberg zgold@bluesata.com wrote:
All,
So it seems to be almost every other day now when somebody mentions how broken the appdb rating system is. This "broken-ness" has had several high level discussions on Wine-Devel. There have also been many, many proposed solutions to this problem.
I do not wish to discuss these problems or these solutions in this thread. It is time something was done about this. By done I mean agreed upon (probably the harder part) then implemented in code and published to the winehq server(s).
I'de like to arrange an IRC discussion for all those interested in finalizing an attack plan, and then assigning the responsibility of doing it to somebody who is capable and has the time (perhaps myself).
Lets (arbitrarily) set a time for the chat at:
When: 11PM UTC (6PM EST) 2/6/09 (Friday) Where: #winehackers
If this arbitrary time seems to be poorly chosen for a large number of people then please propose an alternative well in advance so all are aware of it.
--Zach
Thanks for taking the initiative! Count me in.