AppDB Upgrades

Code

A fork of WineHQ AppDB has been setup at repo.or.cz:

http://repo.or.cz/w/NewAppDB.git

email Zach if you're interested in contributing and you can be added to the project.

Discussion Outcomes

Problems

• Ratings often inflated and inconsistent
• Info can be hard to find
• once found is often hard to understand
• bugzilla: "Show apps affected by this bug" link is hard to see/use (takes two clicks to get to app)

Goals

• Provide users with key factoids
• Information must be accurate and consistent
• Also provide enough in depth information for apps which need it, in an easy manner.

Mockups

• Here is a basic design for a tabbed AppDB page that gives the user what they're looking for right on the front page: appDB-mockup.odt

  • I don't understand that document, at all.

    • The design is for something that allows a user, at a glance, to see what other users thought of it before having to read test results and howtos and such. A moderator ("maintainer") reads the test results for them, picks the appropriate rating, we give the user that rating and what it means, then right below that we have a very easy to fill out one-click survey to ask people the two main questions they have ("will it work" and "how much effort do I have to put in"). This makes it very easy to contribute that form of feedback, and we'll get many more results than we do full test reports.

mock survey based on irc discussion

(May allow upload / pushing up of pieces of registry which specify if dlls are overriden or not) (Almost all questions should have an 'other' field)

Section 1 : Installation

• Please chose the option most accurately describing your install experience
  • Did not use installer or copied from win. install
  • Worked flawlessly with no effort, same as windows
  • Had issues, needed to use override (Please list overrides / upload reg with util)
  • Requires custom patch from bugzilla or elsewhere (if so, show text entry for URL with "If available, please provide url" prompt)
  • Never got install to work, broken.
  • the application didn't have an installer
• If you used winetricks, please list the commands you used
  • text input
• Is this installed from a windows partition? (If so, spit out warning)
• Did you have to do anything else, asides from winetricks commands, to run the app?
  • Yes or No
  • If Yes: text input AND/OR list of checkboxes for common problems. Additional text box option for comments.
• Ship (with Wine?) a utility which looks at the registry and ships to the AppDB your current dll override settings. This information, attached to a report, could prove very useful.

Section 2: Running the Application

• Is this a demo or full version?
• Please chose the option most accurately describing runtime experience
  • Worked flawlessly with no effort, same as windows
  • Had issues, needed to use override (Please list overrides / upload reg with util)
  • Requires custom patch from bugzilla or elsewhere (if so, show text entry for URL)
  • Never got the app to run
• If you used winetricks, please list the commands you used OR the dll overrides (and versions) you needed.
  • text input*
• Please list the major features of the application
  • List of text inputs with checkboxes next to them. Checkbox indicates the application worked. Little bit of javascript to have the number of boxes grow as it needs to.
    • Rather than checkbox of working not working maybe more options. Untested, partially working, working, not working, other, etc.
  • Have this list of features shared between reports and allow each report to add new features.
• Did you make any registry changes?
  • List Key, List New Value.
• Did the app run at full speed?

Rating Ideas

• Use a scoring based off of questions and give colors based off of tiers of scores.
• Working / Not working / partially working / etc.
• Two ratings, one for install one for normal. Combined rating which reflects things like if install doesnt work or install does work and app doesn't run its useless.
• "works perfect (without issues) / most things works / lots
  • of things does not work / not usable""

Old Data

• Should mostly be kept side by side until phased out simply by becoming outdated.

Pre-Meeting Info (2/13/09)

What

An IRC Discussion to decide how to properly fix the Wine AppDB Admin System and discuss how ratings should be implemented

Where

#winehackers on irc.freenode.net

When

February 6th, 2009. 4PM EST, 9PM UTC

Agenda

Ratings

Admin System

Some problems with the current admin interface:

1. There is no way to generate a list of apps that don't have maintainers.
2. "Comments management" currently just displays a list of all comments in chronological order, with no indication what app they belong to or if that app has a maintainer. To actually manage comments in unmaintained apps, an admin would have to browse all the entries, one by one.
3. Parent and child comments all have to be deleted individually. There should be an option to delete all child comments along with the parent.
4. Admins cannot edit comments. (This is only an issue in cleaning up the ISO-8859-1 characters.)
5. "Users management" cannot generate a list of all users. Leaving the search field blank and setting "Show first" to "500 results" will produce a list of the first 500 users, with no subsequent pages.
6. Approving new app entries/avoiding duplicates: the form for approving new apps currently displays a list of possible duplicates, along with a "Move Here" link for each. That works fine if the system does find a duplicate, but it often doesn't when it should (a simple typo will prevent it), and there is no way for the admin to directly move the new submission to the proper entry from the approval page if it does not appear on the list generated by the system. A recent example: someone submitted "Microsoft Visual C++ 2008" as a new app. The duplicates list included every app in the AppDB with "Microsoft" in the name, but did not include "Visual C++" in the list. "Visual C++" had to be temporarily renamed to "Microsoft Visual C++" to move the new version submission to the correct app.
7. Combining versions: The "Move child objects" link on version pages brings up a simple list of all versions for all apps (in no apparent order), with no indication of what app they belong to. To be able to find the right version to move to, one has to temporarily rename the target version to something distinctive before attempting the move.
8. Email notices: the choice currently is either all off or all on. For admins, having notices "on" means receiving an email notice every time anyone enters anything in the AppDB--hundreds per day. It would be nice to be able to opt to receive some types of messages (e.g., notices for apps for which the admin is also a maintainer) without having to receive them all.
9. No way to search only the AppDB by keyword: the "Browse apps" page does not have a keyword filter; the Google search box searches all of WineHQ. (Also affects ordinary users.)

Related Bugs (some of which may need to be closed as fixed)

• would like to be able to view screenshots by application category

• AppDB note editor whitespace generation

• Improve appdb screenshot handling

• Links from http://appdb.winehq.org/ always open in the same window

• Ability to move child versions between apps

• New screenshot notification is misleading

• Editing a howto doesn't work properly

• Add Wine version to the screen-shot

• Appdb's google search fails on Konqueror 4.0.x

• Flat comment mode does not display any replies

• Appdb doesn't remember logins

• Linking a bug to an application instead of a version

• AppDB votes system is being cheated

• create_table option forcreate_user.sql is incorrect

• AppDB should allow more customizable e-mail options

• Allow deleting multiple comments

• Test Data reports should include wine stable versions (i.e. 1.0, 1.2...) and not just recent betas

• Cannot change my E-mail address

• Add language field for applications

• Unclear how to link a bug report to a still queued application / version

• AppDB doesn't confirm test data being aсcepted for review

• Missing categorising by video driver version in AppDB

• Add field for changes to default winecfg settings to AppDB test report form

• processing queued test data: misleading description

• AppDB admin backend enhancements

• Main Body to wide

• broken encoding or database conversion regarding Umlaut

• broken encoding in AppDB PM E-Mails

Pre-Meeting Info (2/6/09)

What

An IRC Discussion to decide how to properly fix the Wine AppDB rating system

Where

#winehackers on irc.freenode.net

When

February 6th, 2009. 6PM EST, 11PM UTC

Agenda

• Establish goal of the upgrade. Think of a user scenario. What information does a normal user want/need from an AppDB Page that they don't reliably get now?
  • Have multiple "ratings" for each major aspect of running an App in Wine. (Installation vs. running vs. missing features, desktop integration etc.) See "Two Ratings" Idea below.
• Discuss methods of AppDB Rating (Possible to have different methods for different purposes or kinds of users (maintainer, tester, random commenter)
  • Survey / Questionnaire (Balance detail / ease of use in coming up with question list)

    • Survey in sections

      • Section on installation / uninstall
      • Section on running
      • Section on missing / broken features
      • (Benefits of this is that we can generate a "score" or rating for each section, giving users more high level information)
• The AppDB interface needs to be adjusted or even redesigned to accommodate the new method of input. Ideally the interface itself can increase the accuracy of the report by providing hints.
  • How will collected information be displayed? Use the maintainer? Talley based on number of surveys? Show all reports? Tabbed interface? etc.
  • Need design ideas. Any good role models here?
• What to do with old ratings?
  • Could keep old rating levels but have them be data driven based on test results instead of set by the maintainer. This was the original plan for phasing out the existing ratings set by the application maintainer. Algorithm could look for a ratio of "working" test results vs. not working test results that exceeds a particular number. -- Chris M.
• Who will implement. What QA to do?
• What parts do we require registration for? Test results, surveys, simple voting?

A Section Concept, High Level

• Did you test every feature of the application?
• Were any tested features anything less than perfect?
• Were any tested features somewhat-off but usable?
• Did you have to consult google or the appdb to make things work?
• How much time did you put in to make the application work?
• Were you able to use the app for its intended purpose for more than a demo? (Aka play through multiple levels of a game, do an actual workflow in a editor application etc.)

A Survey Concept, In Detail

Which additional runtime[s] must be installed first? (Or perhaps what winetricks commands did you run?)

• (please report the used version, when possible)

[CK: could Wine add a button to the title bar to open the App DB for this app, and optionally post this info in a new/existing survey for this user]

• dotnet (.NET)
  • not needed
  • mono, version *1.x / *2.x / xxxx
  • dotnet, version *1.1 / *2.0 / *3.0 / *3.5 / xxxxx
• Visual Basic Runtime
  • not needed
  • version *3 / *4 / *5 / *6 / xxxx
• Visual C/C++ Runtime
  • not needed
  • version 4 (mfc40)
  • version 6 (mfc42, msvcp60, msvcrt)
  • version 2002 (mfc70, msvcp70, msvcr70)
  • version 2003 (mfc71, msvcp71, msvcr71)
  • version 2005 (mfc80, msvcp80, msvcr80)
  • version 2008 (mfc90, msvcp90, msvcr90)
• Windows Scripting Host
  • not needed
  • version *5.6
• Flash player
  • not needed
  • version *9 / 10
• other runtimes
  • not needed
  • xxxxxxxx textfield here

Install the application

• The reported Windows version
  • unchanged * changed to: xxxxxxx
• dll overrides
  • unchanged * using native dlls for: xxxxxxx
• copy protection
  • unchanged * using a patched file
• Not needed
• Installs without Problems
• Installs reached the end, but has failures
• failed to Install
• this is an old appdb result (read only)
  • Describe the failures here:
  • missing dlls / applications
  • crash on unimplemented function: dll, function
  • xxxxxxxx textfield here

Run the application / game

• The reported Windows version
  • unchanged * changed to: xxxxxxx
• dll overrides
  • unchanged * using native dlls for: xxxxxxxx textfield here
• copy protection
  • unchanged * using a patched file
• used graphics mode:
  • window (unchanged) *virtual desktop
  • depth: 32/24 / 16/15 / 8
  • xxxxxxxx textfield here

[CK: Could you add to the right of the rating a comment box to allow a textual explanation of the result. In aggregate, the rating summary is fine, but then you could get more information/understanding if needed]

• starting the application: *perfect / *good / *partial / *failed
• play the intro: *n/a / *perfect / *good / *partial / *failed
• using the menus: *n/a / *untested / *perfect / *good / *partial / *failed
• start a new document/game: *n/a / *untested / *perfect / *good / *partial / *failed
• load / save a document/game: *n/a / *untested / *perfect / *good / *partial / *failed
• edit document / play the game: *n/a / *untested / *perfect / *good / *partial / *failed
• print document: *n/a / *untested / *perfect / *good / *partial / *failed
• video/graphics output: *n/a / *untested / *perfect / *good / *partial / *failed
• sound output: *n/a / *untested / *perfect / *good / *partial / *failed
• using the internet: *n/a / *untested / *perfect / *good / *partial / *failed
• multiplayer mode works: *n/a / *untested / *perfect / *good / *partial / *failed
• using the help: *n/a / *untested / *perfect / *good / *partial / *failed
• this is an old appdb result (read only)

[CK: could you perhaps make "other things work/don't work" match above. Have user click to add new test case, a text description, a rating, and a text description of the result. Test cases would then be required for future testing... to build a base of test cases. Moderators would help 'refine' test cases over time]

• Other things, that works:
  • xxxxxxxx textfield here
  Describe the failures here:
  • missing dlls / applications
  • crash on unimplemented function: xxxxxxxx textfield here
  • xxxxxxxx textfield here

Uninstall the application / game

• The reported Windows version
  • unchanged * changed to: xxxxxxx
• untested
• n/a
• uninstall without problems
• uninstall reached the end, but has problems
• failed to uninstall
  • Describe the failures here:
  • crash on unimplemented function: dll, function
  • xxxxxxxx textfield here

Other Ideas

Two ratings Idea

• Another idea might be to simply have two ratings - one for how well the application runs, and one for how hard it is to install it. (winetricks, custom patch, registry setting, etc.)

• Originally saw the idea on post #9 of: http://ubuntuforums.org/showthread.php?t=1058780

Garbage Ratings

• If linking to Bugzilla bugs is going to be kept in the new AppDB, and if an app is considered garbage or crashes, could it be required to file a bug in bugzilla and link to it?

[DR: Bug-Report only as a suggestion (with a big "please", but not a requirement. That would prevent some people to send the test results]

Making Comments and Fixes relative to Test Results

• With the actual AppDB structure it's impossible to define if the fixes and the comments below the results are still valid, if they apply at different wine and winetricks releases and whatnot, creating confusion upon the users regarding the right approach to install and start the application.It should be clear which comments relate to different tests.
• It's not quite clear how to point a possible regression between wine versions without requiring a link to a bug in Bugzilla.
• At the same time it should be made clear if what the user suggests is just a comment,a suggestion or a necessary fix to run the application.More than often the "fixes" (green), "Notes"(blue), and "Warning" (red) labels are just misunderstood , not enforced or ignored.
• This is especially true for the games section where a lot of comments are often contradictory even when relative to the same game and wine version.

2/6/09 Meeting Transcript

(Times in MST, UTC - 6)

(16:58:01) zgold (n=zgoldber@dsl092-054-196.phl1.dsl.speakeasy.net) joined #winehackers
(16:58:25) Craig73 (n=craig@CPE001ee5377c19-CM0018682d06e8.cpe.net.cable.rogers.com) joined #winehackers
(16:58:35) <YokoZar> Hello all
(16:59:16) <zgold> YokoZar: hi
(16:59:41) r3set quit ("/quit per aspera ad astra ;)")
(16:59:50) <YokoZar> I was thinking we should start a Gobby session
(17:00:19) <zgold> We should stick with IRC for now because thats what people are expecting.
(17:01:20) <YokoZar> We don't all need to take notes
(17:01:54) <YokoZar> Hell Gobby would work even if only one or two of us was doing it (and the gobby document we'd be editing would be the wiki page)
(17:02:41) <zgold> if you want to start a session go ahead
(17:03:33) arethusa (n=arethusa@unaffiliated/scguy318) joined #winehackers
(17:04:09) <zgold> So, I guess its time to start ish.  Who here is for appdb?
(17:04:24) <Craig73> I am
(17:04:30) <YokoZar> I as well
(17:04:52) <austin987> howdy
(17:05:13) <YokoZar> I have a whole lot of little notes to talk about scrawled on a notebook next to me that I haven't had time to type out yet, so forgive me if I mention something seemingly out of nowhere
(17:05:26) *jeffz waves
(17:05:28) dimesio (n=dimesio@user-11fb192.dsl.mindspring.com) joined #winehackers
(17:05:30) <zgold> Hm, I guess we'll wait 2 more mins.  See if anybody else shows up.  If not 4 is good for me
(17:05:53) <jklehm> I'm listening in but my participation will be limited :)
(17:06:05) <austin987> dimesio, is I'm sure interested
(17:06:19) <dimesio> Yes.
(17:06:50) *riclas wrists
(17:07:00) <riclas> can't participate, have an exam in the morning :(
(17:07:08) <zgold> Alrighty, let us begin.
(17:07:16) <zgold> So step 1: Whats wrong now?
(17:07:30) <dimesio> ratings inflation
(17:07:43) <YokoZar> It's hard to find what you're looking for as a user
(17:07:51) <zgold> YokoZar: as in you can't find the app?
(17:07:53) <jklehm> inconsistent measurement
(17:08:01) <zgold> YokoZar: or you can't find the exact steps to make things work
(17:08:26) mellery2 (n=mike@pool-151-199-91-177.roa.east.verizon.net) joined #winehackers
(17:08:43) <YokoZar> zgold: finding the app is reasonable enough (though migrating through the version numbers is sometimes annoying), but once you get there it's not obvious if an application will work or if it won't work or if you'll need to do a complicated howto
(17:08:48) <Craig73> ratings inaccurate - was silver for many releases an then one review makes it bronze (a review that tests nothing)
(17:09:36) <zgold> Alright so i've got that the given color ratings are near meaningless and worse dangerous because they over inflate.  I've also got that the HOWTOs are sometimes complicated and inconsistent
(17:09:46) <austin987> yes and yes
(17:09:55) <YokoZar> The amount of fiddling you need to do isn't obvious, nor how annoying it is, and we don't really have a way of measuring this
(17:09:57) <zgold> Anybody wanna throw in a third?
(17:10:07) <winspool> Hi
(17:10:08) <Craig73> I don't think it's that the rating is wrong - it's just inconsistent
(17:10:35) <Craig73> (or incomplete testing so the rating is meaningless)
(17:10:42) <YokoZar> This is more of a style/presentation issue in the interface, but there are too many words and buttons jammed on the main page
(17:10:58) <zgold> Lets stick to the substance of the system, aesthetics for the end
(17:11:01) <YokoZar> For an app I mean -- so much other stuff that it hides what's important ("will it work" and "what do I do")
(17:11:19) <zgold> I like that
(17:11:26) <zgold> Users want to know, several key facts immediately
(17:11:31) <YokoZar> Yeah we'll talk about layout issues and such at the end
(17:11:31) <zgold> and I think the appdb fails those people
(17:11:41) <zgold> if you just want to know "Will it work, and if so how much effort will it take"
(17:11:48) <zgold> Or better yet
(17:11:53) <zgold> Will it work without the command line
(17:12:25) <YokoZar> Agreed, figuring out if it will work out of the box is important
(17:12:52) <zgold> So one goal should be to provide the users that just want key factoids their information (accurately)
(17:13:01) <Craig73> I want to know what works as well / to give me confidence in the rating
(17:13:11) <arethusa> I noticed that some apps that are regularly updated seem to have juicy bits of information scattered in obsoleted versions, so maybe version cleanup?
(17:13:26) <arethusa> *get new version releases, e.g MMOs
(17:13:37) <zgold> arethusa: that goes under the inconsistent / hard to find info and howtos i think
(17:13:46) vadi2 (n=vadi@H116.C195.cci.switchworks.net) joined #winehackers
(17:13:47) <YokoZar> zgold: and also to collect that information accurately
(17:14:00) <austin987> 'accurately' is the root of the problem IMHO
(17:14:37) <YokoZar> So, clearly the current testing reports and comments are not enough
(17:14:40) <zgold> (Updating wiki with the list of goals we generated)
(17:15:05) <YokoZar> Scanning through all comments is something no one does (or should do, since most are obsolete)
(17:15:09) <austin987> it's not that they're not enough, but like dimesio said, overinflation/inconsistent
(17:15:16) <YokoZar> zgold: by the way my sobby server isn't working so we'll just use the wiki
(17:15:30) <jeffz> YokoZar: maintainers usually scan the comments and convert any useful info into notes/howtos
(17:15:35) <austin987> e.g., if I use a dozen winetricks overrides, is the app now platinum?
(17:15:45) <YokoZar> jeffz: Yes, maintainers are doing a relatively good job of reading comments
(17:15:48) <austin987> jeffz, what about apps with no maintainer?
(17:16:10) <jeffz> austin987: yes, that's a short coming, old comments with bad information and hints from 2006 or so :P
(17:16:13) <zgold> btw if anybody is not OK with a transcript of this chat going on the wiki please pm me
(17:16:30) <zgold> austin987: is "outdated" information a common problem?
(17:16:53) <Craig73> or even a version or two ago... I was wondering if we should aggregate information under "wine 1.0" "wine 1.2"
(17:17:05) <Craig73> some people might be using LTS distributions with an old wine version
(17:17:26) <austin987> zgold, I don't use the AppDB much, honestly
(17:17:28) <zgold> Craig73: I'de go almost one step further and agregate by Ubuntu version, almost.
(17:17:33) <austin987> unless linked in from bugzilla
(17:17:43) <arethusa> we do advise those folks to use dev Wine if possible
(17:17:49) <zgold> Or aggregate by Wine version and indicate which version is in which major distro release
(17:17:57) <austin987> but, for example, I maintain Morrowind, since about 4 months ago
(17:18:19) <austin987> and when I looked in the comments, there were guides from 2-3 years ago, talking about 0.9.30 :-{
(17:18:25) <zgold> Alright
(17:18:29) <zgold> So we have a list of problems
(17:18:42) <YokoZar> It's a bit weird with test results - on the one hand they're not detailed enough to produce reliable ratings for a maintainer to pick, on the other hand they're too detailed for an average user to answer those two questions ("will it work" and "what do I do")
(17:18:57) <zgold> I think we all have a sense of the major issues.  Item number two is the broad category of what kind of system do we want to impleent which fixes these problems
(17:19:22) <YokoZar> Wave the magic ajax/web 2.0 wand at it
(17:19:40) <austin987> ahh, the big question now
(17:19:43) <YokoZar> I believe that was the magic fix discussed at wineconf ;)
(17:19:43) <Craig73> I was hoping to split test results individually... so they can be re-used by other "testers", compute the rating, etc.
(17:19:45) <zgold> The most common idea is to make a rating submission some form of survey, rather than a free-form text box
(17:20:26) <austin987> zgold, I'm a big fan of that as well
(17:20:33) <zgold> Anybody have any problems with the survey method?  If so speak now
(17:20:35) <Craig73> zgold: sounds good.  I was wondering how much info could be sent from Wine automatically / make it easier
(17:20:43) <YokoZar> What about making the maintainer ratings more important than the tester ratings?
(17:20:54) <austin987> automatically was discussed at wineconf, agreed on, then shot down on -devel :-/
(17:21:07) <jeffz> YokoZar: I think there's some logic to prefer the maintainers rating over users rating already
(17:21:19) <Craig73> OK / perhaps design with it in mind for future enhancement
(17:21:20) <zgold> YokoZar: I agree with jeffz on this one
(17:21:24) <zgold> Craig73: agreed.
(17:21:28) <YokoZar> It seems like reading test reports and then picking an accurate rating is something we could do easily, especially if we had admin-type users who could rate many applications
(17:21:45) <YokoZar> zgold: jeffz: What if we split the tester ratings from the maintainer rating in the interface?  Like if tester ratings were on a separate tab.  As it is now they're just as visible on the page
(17:21:49) <zgold> That doesn't really solve the problem of getting consistent info
(17:22:07) <Craig73> it also complicates things / and creates a bottle neck
(17:22:10) <zgold> YokoZar: Lets step away from the flat out "ratings" idea for now.  This could change.
(17:22:12) <YokoZar> You mean info in the test reports, or information about the ratings?
(17:22:35) <austin987> YokoZar, anyone can add results, problem is (at least for me), time consuming
(17:22:38) <zgold> YokoZar: consistent anything.  As long as its free form even if its free-form from moderators we don't solve the problem of generating a "it works" or "it doesn't work" for =users
(17:22:46) <austin987> I test dozens of apps a day, but don't take time to submit to AppDB
(17:22:55) <zgold> austin987: shame on you =P
(17:22:59) <Craig73> no free form...
(17:23:12) <austin987> zgold, somewhat...but I'm looking for bugs, and once I've found that, I don't test further
(17:23:16) <YokoZar> At some level we are going to have to rely on maintainers since there will be conflicting test results no matter what (different systems, etc)
(17:23:18) <Craig73> have each test case entered as a separate line item, add a rating, and a comment of test result
(17:23:24) vadi2 (n=vadi@H116.C195.cci.switchworks.net) left #winehackers
(17:23:25) <austin987> having a few people just test apps all day would be good
(17:23:44) <winspool> I collected some questions, that the tester yould easy answer with a checkbox.
(17:23:55) <zgold> Right, so I agree we should think about the survey idea
(17:23:57) <Subdino> If I may intrude with random ideas: 1) Provide a way to rate an app by the number of enabled quirks (bronze with pure wine, silver with dll overrides, etc). 2) Provide some checklists for "what has been tested", with probably a game-wide checklist describing all features and a 3-choice "working-not tested-not working".
(17:24:01) <zgold> Anybody against using a survey?
(17:24:33) <YokoZar> zgold: I'm slightly worried it will be harder to submit a test result.  So austin987 will be even less likely to put his tests on appdb
(17:24:51) <Craig73> difficulty is in UI design no?
(17:24:59) <winspool> An important thing would be a selection, what is a main feature for an app/game
(17:24:59) <zgold> YokoZar: I'm invisioning something like 10 or 15 checkbox or radio button questions followed by a text box or two for comments
(17:25:01) <austin987> YokoZar, actually, I'd find a surver easier
(17:25:04) <austin987> survey*
(17:25:14) <YokoZar> Craig73: Well, collecting 90 million bits of information would be difficult no matter how you did the interface ;)
(17:25:16) <dimesio> I prefer a survey too.
(17:25:22) <austin987> because I often only test installers, or specific functions
(17:25:23) <Craig73> true
(17:25:49) <austin987> so if I could find the app, check off installer works, runs after native riched20, I'd be happy
(17:25:50) <Craig73> what I had been suggesting / fwiw / is the ZGOLDs survey, the basic questions to get people going
(17:26:04) <Craig73> Then a box to type in a test case and how it worked
(17:26:20) <austin987> now if we could integrate appdb/bugzilla to where I could click the AppDB entry from bugzilla, you'd get a lot more results from me :-)
(17:26:20) <winspool> Working Internet for an ftp client is required, but for office only a nice to have improvement
(17:26:20) <Craig73> which when typing it in would create another text box below (empty)for the next test case
(17:26:21) <zgold> alright so lets think about what we need in the survey itself
(17:26:38) <YokoZar> austin987: Can't you do that already?  Or almost already?
(17:26:40) <jeffz> I don't know how we can stop people from using unclean prefixes, eg if they've used winetricks to install a bunch of stuff for an app other than the one they're rating.
(17:26:41) <Craig73> then the survey is extendable and specific to the app
(17:26:58) <Griswold> I was afk
(17:26:59) <zgold> jeffz: add a disclaimer on top to only submit results from a clean wineprefix?
(17:27:08) <zgold> Griswold: good luck reading log =P
(17:27:13) <Griswold> zgold, Just did :P
(17:27:15) <YokoZar> jeffz: we could ask them if they tried it clean in the survey
(17:27:23) <Subdino> jeffz: what about generating dumps of configuration, and something along the lines of ldd ?
(17:27:24) <austin987> YokoZar, well, I mean more of an appdb specific link somewhere
(17:27:41) <austin987> you can currently paste it in the comments, but not a simple/obvious way to do it
(17:27:43) <jeffz> Subdino: it's a bit of a hassle to search the entire prefix
(17:27:55) <austin987> e.g., url box, keyword box, appdb link box
(17:28:01) <jeffz> zgold, YokoZar: I guess the least we can do is ask them to do the right thing.
(17:28:08) <zgold> Next agenda topic to be agreed upon is whether or not the survey should be split into discint sections.  E.g. section for install, run, iuninstall etc.
(17:28:28) <zgold> And what those sections should be
(17:28:34) <YokoZar> austin987: There's a "show apps affected by this bug" button isn't there?  See up top: http://bugs.winehq.org/show_bug.cgi?id=13675
(17:28:36) <buggbot> Bug 13675: enhancement, P2, ---, wine-bugs@winehq.org, UNCONFIRMED, emule upnp not working
(17:28:37) <zgold> I think this should be based off the key areas of information users need
(17:28:49) <vpovirk> and I just remembered about the meeting :/
(17:29:00) <zgold> vpovirk: no problem =P
(17:29:08) <austin987> ahh, well crap
(17:29:13) <Griswold> At least difficulty / what is needed for install and how well the prog. runs / what is needed to make it urn
(17:29:14) <Griswold> *run
(17:29:15) <austin987> never noticed it, so that proves my point
(17:29:32) <vpovirk> should I read backlog?
(17:29:41) <Griswold> Wine uninstaller doesn't work well right now, so not sure we should have that until we fix the uninstaller program
(17:29:44) <zgold> vpovirk: can't hurt, griswold would know how useful that is
(17:30:10) <YokoZar> zgold: People who don't test installation might still have good test results (eg, they installed fine with earlier wine version) -- in fact I suspect this is how most tests are generated
(17:30:12) <austin987> Griswold, agreed
(17:30:15) <zgold> Griswold: so you suggest something like "How well did the install work" section, "What did you have to do to get install work" section, and the same two kinds of things for running
(17:30:35) <zgold> YokoZar: so an option in the install section can be "had to work around installing as it didn't work"
(17:30:35) <winspool> I'm for a seperation on install and run
(17:30:39) <zgold> or something tothat affect
(17:30:40) <Griswold> I suggest radio buttons under a section for "Install"
(17:30:47) <jeffz> if people don't get accustomed to testing installing, then there's a risk that regressions go uncaught
(17:30:57) <Subdino> jeffz: at least dumping winecfg settings for a given app would be a start to detect overridden dlls (although not perfect)
(17:31:04) <Griswold> And then the radio buttons:  installed perfectly, had to use winetricks, ..., other _____
(17:31:16) <austin987> jeffz, I like the installer section, but if it's 'perfect', disable the 'what did you have to do'
(17:31:18) <Griswold> Check boxes might be better
(17:31:23) <zgold> Do we think having users dump registry (hence winecfg) and upload that to AppDB would be useful?
(17:31:44) <YokoZar> zgold: more like "already had it installed from previous wine version"
(17:31:58) <zgold> We could automatically find out what dlls are overriden, and if other apps are installed etc.
(17:32:02) <Craig73> but if most questions have "untested" we should get a good idea of how well the install is working / because we can look at cases where it was tested
(17:32:22) <YokoZar> jeffz: only installer regressions.  We shouldn't throw out the test report if that's too much for them, just encourage them to do the full test
(17:32:27) <winspool> We should not expicit mention winetricks.
(17:32:33) <zgold> winspool: why?
(17:32:34) <jeffz> YokoZar: fair enough
(17:32:37) <austin987> zgold, I don't think registry info would be useful
(17:32:43) <zgold> winspool: it seems almost every report says "use wintricks for  blah" as step 1 or 2
(17:32:55) <austin987> perhaps winecfg settings, e.g., a box to list changes from default, but not full registry
(17:33:08) <winspool> Nothing prevent us from extending "Add/Remove Apps" to download and Install Mono or other runtimes.
(17:33:11) <zgold> austin987: so a custom app to upload a diff of your reg vs. default
(17:33:15) <austin987> winspool, I disagree, winetricks is very useful, especially for mfc42, etc.
(17:33:21) <Griswold> Maybe "perfect," "already installed," "failed to install," "had to do..."
(17:33:30) <zgold> Griswold: I like.. will record
(17:33:39) <austin987> zgold, possibly, but that would also give you a lot of crap from the installer...perhaps a diff on the wine sections only
(17:33:58) <zgold> austin987: would this be better than asking reporter to manually say which overrides they did?
(17:34:05) <Craig73> I was under the impression winetricks was necessary in some cases / it's the overriding dll's that should reduce the rating of an app
(17:34:11) <Griswold> Then under "had to do..." put checkboxes for winetricks, registry alterations, winecfg change, patch wine, patch game/app, other: ____
(17:34:13) <zgold> austin987: we could provide some little util which uploads this info to your account on appdb
(17:34:17) <YokoZar> zgold: if it's more accurate and less bothersome, yeah.  Especially when a tool made those overrides for them.
(17:34:26) <winspool> It should not look like winetricks is a requirement to use.
(17:34:26) riclas quit ("CyberScript - President Bush uses CyberScript. Shouldn't you? (www.cyberscript.org)")
(17:34:45) <austin987> zgold, I don't know to be honest, I'll have to think about that one
(17:34:46) <Griswold> dll override should be a checkbox too
(17:34:47) <YokoZar> Craig73: you'd be right, especially for things like mono / .net required, since we aren't bundling those (yet?)
(17:34:51) <Subdino> austin987: i would not suggest dumping the entire registry, but winecfg subtree, and *maybe* some special keys to know if user installed some known programs (like from winetricks)
(17:35:00) <zgold> winspool: theoretically yes.  Wine should never need winetricks.  But if our goal is to give the user the best inforation as easily as possible it might be worthwhile
(17:35:01) <winspool> We can of course mention the winetricks option for a runtime.
(17:35:03) <austin987> austin987, that'
(17:35:12) <austin987> Subdino, that's what I was saying ;-)
(17:35:19) <vpovirk> I think a lot of people are not going to download an app or upload their registry to send reports
(17:35:28) <Subdino> austin987: I'm too slow -_-'
(17:35:29) <zgold> vpovirk: if the app comes bundled with wine?
(17:35:34) <zgold> or wine --uploadregtoappdb
(17:35:38) <winspool> I did 2 seperate ways for VB and VC runtime.
(17:35:46) <Craig73> the separate utility that uploads winetricks info could also upload the apps wine configuration info
(17:35:48) <zgold> julliard: would that be accepted? :)
(17:35:51) <Griswold> What do you guys think of the checkboxes under "had to do..." idea?
(17:36:01) <zgold> Griswold: I like it.  Wrote down your options
(17:36:05) <austin987> Griswold, I like it, just what I was thinking
(17:36:10) <Griswold> Good.  :)
(17:36:14) <dimesio> I like it.
(17:36:30) <zgold> And an almost identicle set of options for running the app itself
(17:36:42) <Craig73> agreed
(17:36:50) <zgold> So... let me mock up a survey on the wiki
(17:37:01) <austin987> zgold, I don't think the app belongs in wine itself, and it probably would be a pain to get users to do it...wasted effort
(17:37:03) <Craig73> (wasn't there one there already)
(17:37:22) <zgold> Craig73: there were some mockups, but no mockups based on thisconversation
(17:37:42) <Craig73> zgold:a starting point though
(17:37:44) <zgold> I'de like to generate a survey and a list of rules for generating ratings from this eeting
(17:37:49) <zgold> meeting*
(17:38:04) <Craig73> OK
(17:38:34) <jeffz> If any data is going to be collected, I don't think you'd want to upload the entire registry, perhaps only a script which tests certain registry keys and searches the prefix for native dlls to provide a summary.
(17:39:24) <zgold> jeffz: agreed
(17:39:25) <Griswold> Do we even still want to have the rating colors?
(17:39:28) <zgold> question 1 on the wiki
(17:39:33) <Craig73> could it also provide the other info suggested in the wiki - window / colour depth, OS simulated, etc.
(17:39:35) <Griswold> plat, gold, etc.
(17:39:41) <zgold> Griswold: Maybe... a discussion for 30 minutes from now =P
(17:39:46) <Griswold> Ahh, ok
(17:39:56) <YokoZar> I agree with keeping in mind that any sort of script we ask the tester to run should be as painless as possible so we don't scare them away entirely.
(17:39:58) <Craig73> [I might not be here in 30... so my vote is yes]
(17:40:29) <Griswold> Any script should be optional also
(17:40:42) <zgold> So, check the wiki
(17:40:45) <Griswold> Or just automate what they would input
(17:40:45) <Craig73> [sorry my vote yes was for Griswold]
(17:40:47) <zgold> I've got 2 and a half questions there
(17:40:48) <Griswold> Ok, checking wiki
(17:40:52) <zgold> What else do we want to ask about installtie
(17:40:59) <zgold> Menu entries were added?
(17:41:30) <zgold> Do we want a question about using a CD or an ISO?
(17:41:33) <Griswold> Some programs don't have installers
(17:41:35) <zgold> or a .exe ?
(17:41:36) <vpovirk> menu entries almost always work
(17:41:44) <Griswold> ie. just unzip into a dir or the like
(17:41:50) <vpovirk> I doubt it's worth asking about them specifically
(17:41:52) <zgold> Griswold: added that option
(17:41:55) <Griswold> ok
(17:42:02) <zgold> vpovirk: for which question
(17:42:09) <vpovirk> menu entries created
(17:42:13) <zgold> mkay
(17:42:25) <vpovirk> where am I checking the wiki?
(17:42:25) <zgold> Thoughts on asking about using an ISO or a CD or .exe ?
(17:42:31) <jeffz> I think I've seen more cases of someone trying to use an iso that has failed vs using the original media.
(17:42:32) <zgold> AppDBUpgrades
(17:42:32) <Griswold> Lots of people use winetricks to install
(17:42:39) <Griswold> Should add an option for that.
(17:42:45) <austin987> Griswold, to install what?
(17:42:56) <Griswold> "I use the winetricks command: winetricks ____"
(17:42:59) <vpovirk> if we mention iso's, we might imply it's a supported configuration, which it isn't..
(17:43:04) <zgold> Griswold: adding that as we speak
(17:43:13) <zgold> vpovirk: K.  Won't ask that
(17:43:13) <Griswold> Sometimes people have to use it to fix up a registry key or download a dll, install .net, etc.
(17:43:55) <Griswold> For custom patch we should have an optional url field
(17:44:00) <vpovirk> it'd be neat if we could detect that from the prefix, since we're talking about detecting things from the prefix
(17:44:03) <zgold> 5 questions so far
(17:44:20) <zgold> Griswold: will add the url field idea
(17:44:44) <winspool> I suggested a seperate section for the environment (runtimes)
(17:44:45) <austin987> Griswold, winetricks doesn't install many 'real' apps (firefox3/kde3/autohotkey/wmp)
(17:44:45) <zgold> Other instalation questions?
(17:44:56) <Craig73> are these yes no questions? or where are you collecting the info (a box beside the question?)
(17:45:07) <YokoZar> Radio buttons I thought?
(17:45:10) <zgold> winspool: What kind of section title would you give that?
(17:45:19) <Griswold> austin987, Yes, but it installs deps for people (.Net, fixes registry, dcom, etc.)
(17:45:25) <jeffz> zgold: how about asking if the the user is running the app from their windows partition?
(17:45:28) <zgold> Craig73: not detailing form of input yet.  Mostly just 'gisting' the data we want
(17:45:29) <winspool> When we make a table, a column could have the winetricks option for that runtime
(17:45:41) <austin987> right, but that's not the installer for the app itself, that's what I thought you meant
(17:45:49) <zgold> jeffz: I hope not =P... I think that falls under the same category of .iso, don't want to encourage certain behaviors
(17:45:54) puk quit ("Client exiting")
(17:46:01) <austin987> zgold, same for copying the install
(17:46:13) <jeffz> zgold: a good place to tell them not to, since plenty of people still insist on doing so
(17:46:19) <zgold> winspool: Arn't there cases where you need to do different winetricks things for install / runtime
(17:46:24) kumbayo quit (Remote closed the connection)
(17:46:34) <zgold> winspool: aka its a good testcase for wine both the install and in the runtime
(17:46:40) <Griswold> Some applications need special wine commands to install
(17:46:40) <zgold> seperate testcase*
(17:46:56) <Griswold> Like my 6-disc BFME2 install required 'wine eject'
(17:47:03) <Subdino> install-related question when done from cds: "how many cd drives used to install"... I don't know how to put that sentence together
(17:47:20) <Griswold> Might want a textbox somewhere where people can detail what they did if it doesn't obviously fit any other option
(17:47:21) <vpovirk> there probably are cases like that, but I don't think it's reasonable to ask users to guess which process they needed a particular winetricks command for
(17:47:28) <Griswold> Or an "other thoughts" box
(17:47:32) <winspool> zgold: One runtime section to reduce the number of questions.
(17:47:41) <Craig73> could the winetricks command [text input] be broken down (parsed) for individual packages
(17:47:45) <Subdino> basicaly, I cannot install star wars force commander (last time I tried) because I cannot change the cd. I bet some game can only read the second disk if there is a second drive
(17:48:00) <Griswold> Subdino, Try using 'wine eject' next time
(17:48:05) <Subdino> "cannot change the cd" as in "device in use"
(17:48:05) <jeffz> Subdino: you need to use wine eject in those situations
(17:48:12) <Subdino> gah
(17:48:15) <Subdino> nevermind then
(17:48:25) <Griswold> Although this should be more transparent to users
(17:48:43) <winspool> Or pre-select runtime-options for "run" with the options from "install",
(17:48:44) <zgold> please refresh wiki page
(17:48:48) <vpovirk> you might need an override AND a patch
(17:48:50) <jklehm> slightly behind topic pace and I do have to go now,  but to address why there isn't more appdb bugzilla integration:  The main reason is the lack of a unified login system/session for wine. It's been on the appdb todo list forever  http://wiki.winehq.org/AppdbInfo
(17:48:50) <Griswold> Maybe a sub-field under "install by CD" for "used wine eject"
(17:48:51) <Griswold> ?
(17:49:00) <winspool> but we should not require to many questions
(17:49:01) <Craig73> Griswold: perhaps a helper toolbar running during installs - with the eject button
(17:49:02) <vpovirk> and maybe something else
(17:49:14) <Griswold> Craig73, Good idea.
(17:49:16) <jeffz> Griswold: I've thought about making a system tray app with a 'eject cdrom' option on its menu
(17:49:26) <zgold> Focus!  Survey
(17:49:34) <austin987> Craig73, parsing it wouldn't help, you can override by copying dlls from windows, etc.
(17:49:39) <vpovirk> so we maybe need something like "Had issues, needed to" and a series of checkboxes?
(17:49:47) <zgold> vpovirk: I like
(17:50:13) jklehm (n=john@CPE-69-23-199-212.wi.res.rr.com) left #winehackers
(17:50:40) <Craig73> Austin987: the question was about winetricks / I also thought it might be useful to see all the times we override a specific DLL (maybe???)
(17:50:55) <austin987> zgold, along with a textbox of course for unmentioned issues
(17:51:17) <austin987> Craig73, that info can be gotten from the registry...winetricks doesn't record where/what it did
(17:51:19) <zgold> austin987: noted.
(17:51:22) <vpovirk> being able to see what dll's we're overriding and what runtimes we're installing most often would be awesome
(17:51:29) <austin987> Dan could extend it to do so, I'm sure, but I don't see the point
(17:51:42) <austin987> it's made to be quick and dirty to get stuff working, no more, no less
(17:51:44) <zgold> vpovirk: The little util idea to push up that info is floating around.  Need Julliard to sign off on it
(17:51:51) <zgold> So, any thing else on install?
(17:51:55) <zgold> If not lets move on to runtime
(17:52:00) <austin987> zgold, one last one
(17:52:02) <zgold> (We're making good progress so far)
(17:52:10) <Craig73> zgold/austin: will the questions be extensible?  So moderators can add/remove more later if necessary
(17:52:16) <austin987> sha1sum/disk size, so we can make sure people are using the same program
(17:52:23) <austin987> or a place to put it, not necessarily required
(17:52:39) <zgold> Craig73: Dunno, depends how its implemented.  I imagine AppDB overlords should be able to modify the survey
(17:52:40) <jeffz> austin987: some programs change so frequently that sha1sum might not be useful
(17:52:55) encoded_ (n=somedude@reactos/tester/encoded) joined #winehackers
(17:52:57) <austin987> jeffz, agreed, but that's why we have different versions ;-)
(17:53:00) <Griswold> Make sure there is an "other"-type field for people who don't match the standard options
(17:53:10) <austin987> new versions may fix/workaround wine bugs
(17:53:11) encoded quit (Nick collision from services.)
(17:53:17) <Griswold> Actually, the last Q. seems to handle that fine
(17:53:18) encoded_ is now known as encoded
(17:53:23) redblue (i=star@ppp032.109-253-207.mtl.mt.videotron.ca) joined #winehackers
(17:53:42) <zgold> Alrighty, so.  Runtime.
(17:53:45) <Craig73> austin987: perhaps it should be a list, so multiple exe's can apply to the appdb entry
(17:54:04) <zgold> I think we should handle the issue of versioning as a seperate topic post-survey
(17:54:11) <zgold> Lets finish a basic outline of the survey first
(17:54:22) Lemonzest quit ("Ex-Chat")
(17:54:25) <austin987> zgold, sorry, if you don't mind, I'd like people's opinion on the sha1sum/disk size (not necessarily now)
(17:54:26) <YokoZar> zgold: agreed.  Though versioning def. should be discussed :)
(17:54:29) <winspool> Beware, that we must not ask to many question. Think, what you would do, when you get asked to fill a survey.
(17:54:34) <zgold> I agree on both counts.
(17:54:35) <jeffz> if an new version is released weekly, then you're going to end up with a lot of versions
(17:54:43) <YokoZar> winspool: don't want to feel like I have homework
(17:54:44) <zgold> winspool: yeah.  So far we have only 7 questions.
(17:54:49) <Craig73> (OK runtime)
(17:54:52) <zgold> I think 10 ir so is about right
(17:54:55) <winspool> 15 Questions -> ok, i do that.  30 Questions -> too much time for me. someone else can fill it.
(17:55:01) <vpovirk> "Worked flawlessly" should always be the first option btw >_>
(17:55:05) <austin987> jeffz, there are few of those apps (wow comes to mind, but it works well)
(17:55:11) <YokoZar> vpovirk: and marked by default
(17:55:12) <zgold> vpovirk: or last.  We don't want to encourage inflation
(17:55:28) <vpovirk> hmm
(17:55:35) <Craig73> I think the list of basic test works.  Then someone can fill in additional more detailed tests if they want
(17:55:38) <vpovirk> never mind then
(17:55:39) <zgold> So we have to balance the concerns of inflation and ease of use
(17:55:44) <zgold> not sure what the right answer is.
(17:55:55) <zgold> So runtime
(17:56:01) <YokoZar> zgold: It should be very very easy to submit a test result for an app that works flawlessly.  They already know they'll take more time if the app doesn't work since they have to mark what
(17:56:20) <Craig73> yokozar: the problem is people submit - it installed so it's gold
(17:56:21) <zgold> We have the did it work, the winetricks. and the "Test all features".  If they say no do we have them list the things they did do?
(17:56:36) <Subdino> Craig73: +1
(17:56:38) <zgold> Craig73: no notion of gold/silver etc. for now.  We'll probably end up changing that
(17:56:58) <YokoZar> I could be wrong, though I'm not sure people will leave the "everything worked flawlessly" button ticked if things were broken out of laziness
(17:57:14) <Craig73> yes ok - my point more is if we can identify say 5-10 key questions to establish reasonable coverage - or at least what was and was not covered
(17:57:15) <Griswold> zgold, Just have a big textbox for people who did detailed stuff
(17:57:30) <winspool> The install option "copied from Windows" is a no go. Must not be on the same line as "no install required"
(17:57:43) <Craig73> yokozar: I wasn't suggesting that - people just have a different idea of "working" and "testing"
(17:57:57) <YokoZar> winspool: It might not even be worth having "copied from windows" test reports at all
(17:57:57) <zgold> So I have an issue
(17:58:10) <jeffz> winspool: I agree, requiring a copy of windows is a bad dependency
(17:58:15) <zgold> Ultimately we want to generate a "This app works" or not from a survey
(17:58:16) <Craig73> griswold: a big test box is not reusable / nor can it contribute to the ratings
(17:58:27) <zgold> And that will heavily depend on whether or not the user actually tested things in the app
(17:58:29) <Griswold> Like, "I put the first disc in, cd'ed to /media/cdrom0.  wine autorun.exe.  Then did wine eject, put the next disc in until all discs.  After the initial install, I had to go download 1.06 patch from online to get the application working in Wine."
(17:58:34) <zgold> Oh I have a better idea
(17:58:36) <zgold> updating wiki
(17:58:47) <Griswold> ^^ What I did for a recent app, btw
(17:59:23) <Craig73> zgold:  did you see my edits to the original wiki survey?  I suggested (annoying perhaps ;-)  ) to have each test case separate
(17:59:40) <Craig73> then people can add more test cases / and if it was tested and worked or not it contributes to the rating.
(17:59:43) <zgold> Craig73: Refresh wiki.  I added something similar
(17:59:47) <Craig73> if people didn't test it, it doesn't contribute
(17:59:49) <zgold> Craig73: I think its a good iea
(18:00:17) <zgold> The idea is that the AppDB maintains a list of features of any given app
(18:00:22) <vpovirk> so you'd end up with an app-specific checkbox of the things you tested
(18:00:23) <zgold> this list may be generated by people entering reports
(18:00:24) <vpovirk> ?
(18:00:28) <zgold> vpovirk: yes
(18:00:31) mellery2 quit ("Ex-Chat")
(18:00:33) <vpovirk> nice
(18:00:40) <zgold> vpovirk: this allows us to give users very detailed info about how well the app works, and consistent too
(18:00:42) <YokoZar> Craig73: if I get you right, we'd have application-specific test cases then?  And just show those on the survey?  I guess "installation" would be a default test case?
(18:01:09) <zgold> I like this.
(18:01:18) <Craig73> YokoZar: yes, there would be common ones - but for big apps (office, autodesk) there are unique tests that should be performed
(18:01:23) <zgold> So, any other comments for Runtime or new sections?
(18:01:46) <Griswold> Hmm
(18:01:47) <Craig73> zgold: I would like a text box beside each test for comments
(18:01:47) *vpovirk refreshes page
(18:01:59) <zgold> Craig73: good idea
(18:02:15) <vpovirk> hmm, features are really a 3-state
(18:02:20) <Griswold> How about "related bugzilla bugs"?
(18:02:25) <zgold> vpovirk: explain
(18:02:26) <vpovirk> there's a set of features you tested, and then there's a subset that worked
(18:02:33) <Craig73> zgold: this would allow 1) an aggregate view based on test results, and a details view to understand them if they seem off or unclear
(18:02:42) <zgold> vpovirk: untested / works / didn't work ?
(18:02:57) <Subdino> that's the "working-not tested-not working" I suggested earlier, I think
(18:03:01) <Subdino> eh :)
(18:03:08) <vpovirk> how about, first ask for the features you tested (adding if necessary), and then ask for which ones worked?
(18:03:10) <Craig73>  *n/a  / *perfect / *good / *partial / *failed
(18:03:14) <winspool> Instead of "If you used winetricks, please list the commands you used ", i would name that: "If additional runtimes where required, list them here" with predefined options and a text field. Below that the current winetricks question.
(18:03:51) <zgold> winspool: is that too complicated?
(18:03:58) <Craig73> I guess I wanted "not applicable", because if we have a base set of questions, they might not apply to a new app
(18:04:05) <vpovirk> 3-state is confusing, and 5 state is worse :x
(18:04:13) <Craig73> not 5 state - radio buttons
(18:04:25) <vpovirk> it's still 5 choices per feature
(18:04:32) <Subdino> Craig73: I would tend for an opt-in instead
(18:04:33) <vpovirk> and what's the difference between perfect, good, and partial?
(18:04:38) <Griswold> I personally hate filling stuff out when there are way too many choices
(18:04:57) <Griswold> 3 max I think w/ text input when needed and/or at the end works best I think
(18:05:01) <Craig73> well three at a minimum
(18:05:07) <Craig73> OK
(18:05:14) <zgold> Alright this is good.
(18:05:21) <zgold> I think we have a lot of information collected now
(18:05:27) <Craig73> give the moderator a way to remove questions that don't apply
(18:05:32) <zgold> Enough to solve the basic user questions AND give detailed information per feature
(18:05:38) <Griswold> Or four maybe:  works/doesn't work/untested/other: _____
(18:06:08) <Craig73> "other" invites trouble, and can't be calculated
(18:06:08) <Subdino> ...and maybe give the anonymous propose options that applies in their opinion, that would probably help maintainer to test what others expect from the app
(18:06:11) <Griswold> Would be cool to have a field that goes right to appdb devs
(18:06:11) <vpovirk> I really think we need to split it into two stages and have just checkboxes :/
(18:06:22) <Craig73> (calculated against the total score)
(18:06:23) <Subdino> s/give/allow/
(18:06:24) <vpovirk> so much easier to say "I tested these" and then "of those, these worked"
(18:06:24) <Griswold> So people could propose new questions, etc.
(18:06:33) <Griswold> Although that might be better handled on wine-* mailing lists
(18:06:33) <vpovirk> at least, easier to explain
(18:06:45) <zgold> OK
(18:06:46) <zgold> moving on
(18:06:51) <Subdino> vpovirk: +1
(18:06:53) <zgold> Last call for data to collect in the runtime section
(18:07:11) *zgold bangs the auction gabel... going once
(18:07:11) <Craig73> did we get a way to add tests?
(18:07:19) <Griswold> zgold, waiot
(18:07:22) <zgold> Craig73: check the wiki... satisfactory?
(18:07:29) <Griswold> url for patch should be optional
(18:07:33) <winspool> zgold: you must limit the app specific checkboxes to some common options. All data for a test is stored in the database.
(18:07:37) <zgold> Griswold: yep
(18:07:44) <Griswold> If it sounds required, some might give inaccurate info
(18:07:45) <Craig73> sure ok
(18:08:05) <Griswold> "(If so, show text entry for URL)" sounds too forceful
(18:08:12) <zgold> Griswold: Yeah, we'll have some form of option marker... aesthetics =P
(18:08:14) <Craig73> winspool: each app has unique requirements, why limit tests?
(18:08:15) <austin987> a bit late, but remove the copied from windows under installer, or give a big error (thought someone already said this)
(18:08:23) <Griswold> (Please give URL, if possible: ____)
(18:08:23) <zgold> Griswold: text on the wiki is just my quick scribble, not final wording =P
(18:08:46) <Griswold> What about registry keys?
(18:09:06) <zgold> Griswold: the idea is to have an app which can upload to appdb certain keys which are pertinent
(18:09:08) <Griswold> "Needed to alter registry keys to run application"
(18:09:12) <zgold> Oh
(18:09:16) <zgold> Hrm
(18:09:16) <vpovirk> hmm
(18:09:18) <zgold> I like.
(18:09:20) <winspool> zgold: For the runtimes, list some dllnames with the runtime version.
(18:09:34) <zgold> winspool: K
(18:09:40) <vpovirk> only if we can generate a .reg file from the test >_>
(18:09:47) <Craig73> zgold: can you add the per test case comment box to the wiki
(18:10:02) <winspool> Craig73: difficult to search, when you have only free text
(18:10:04) <Subdino> Griswold: would it belong to the "dump a chunk of registry" ?
(18:10:10) <zgold> Craig73: the wiki mentions this
(18:10:32) <Griswold> Not sure if we want to require a registry dump for an appdb entry
(18:10:36) <vpovirk> why do we need per-test-case comments?
(18:10:44) <vpovirk> it seems like it'd add a lot of fields
(18:10:46) <zgold> Griswold: I don't think it'd be required.  That sounds bad.
(18:10:56) <Craig73> zgold: I don't see this ... a text box to enter test case, a rating, and another text box with results?
(18:10:56) <zgold> vpovirk: I think it solves a major user complaint
(18:10:56) <Subdino> not as a requirement, but probably would help filling in data
(18:10:57) <Griswold> zgold, exactly.
(18:11:07) <zgold> vpovirk: that some apps get "Gold" or platinum yet something like "Network Play" doesn't work
(18:11:14) <Craig73> winspool: why would you search?
(18:11:17) <Griswold> Anything like that should be made obvious that it is entirely optional
(18:11:21) <vpovirk> but if it doesn't work that's just "No"
(18:11:25) <vpovirk> you don't need a comment for it
(18:11:29) <zgold> Craig73: Section 2, third bullet, first sub-bullet
(18:11:43) <vpovirk> as long as we have separate test cases, we can very easily indicate that some things work and some do not
(18:11:47) <winspool> The app maintainer (or the one, who added the app) should select the most important features for an app
(18:11:49) <Subdino> Griswold: bet users would prefer click-export-click-import over filling a field
(18:11:54) <Subdino> I*
(18:11:55) <zgold> vpovirk: It could be something like "Network Play" - "No" - "LAN works but WAN won't
(18:12:06) <Griswold> Subdino, Make it optional
(18:12:11) <Griswold> So people can do either/or
(18:12:15) <Subdino> sure
(18:12:17) <vpovirk> hmm
(18:12:18) <zgold> Alright, I think we need to move on
(18:12:27) <Griswold> One more thing
(18:12:29) <vpovirk> I'd think LAN and WAN should be separate
(18:12:30) <winspool> that way, all tests for an app have answers for the same fratures
(18:12:38) <zgold> vpovirk: I agree, that was a bad example =P
(18:12:44) <Griswold> We need a way for users to tell others how to run an application through some obscure method
(18:12:49) <Griswold> Where should that be done?"
(18:12:50) <zgold> Griswold: ?
(18:12:51) <Griswold> Comments?
(18:12:56) <zgold> Griswold: what does that mean
(18:12:59) <Griswold> In the appdb, there are threads/comments
(18:13:04) <vpovirk> well,  no, it points out a problem, namely that you can't cleanly make them separate once someone enters "Network Play" :/
(18:13:13) <Griswold> There is the place for app maintainer to give instructions on how to run the app
(18:13:13) <jeffz> Griswold: the maintainer can create a howto blurb
(18:13:18) <Griswold> ok
(18:13:19) <zgold> vpovirk: maintainers could reword questions for an app I guess
(18:13:23) <Griswold> So just leave it up to the maintainer?
(18:13:35) <vpovirk> although maybe a maintainer would convert it to WAN or deprecate it
(18:13:40) <winspool> Craig73: Printing / Network access / multiplayer / graphical glitches / gfx to bright or dark / ...
(18:13:42) <vpovirk> that's probably good enough
(18:13:55) <dimesio> what about apps without maintainers?
(18:14:05) <zgold> I really don't want to cut off this conversation but theres a lot more to cover.
(18:14:16) <Griswold> Maybe a dynamic option of, "apply to be maintainer of this app" for apps that don't have maintainers?
(18:14:26) <zgold> Griswold: thats not already there?
(18:14:39) <Griswold> Don't know, haven't used AppDB much recently.  :\
(18:14:41) <vpovirk> dimesio: we've covered that; you get a list of the features people have reported on already, and you can add to it
(18:14:55) <zgold> OK.  On to determing ratings.
(18:14:58) <zgold> This one will be hard.
(18:15:08) <vpovirk> but we just got all this great information from the survey
(18:15:09) <Griswold> Do we even still want to have ratings?
(18:15:12) <zgold> We've got this set of information from the report.  How do we want to encapsulate it into something meaningful and portable
(18:15:15) <vpovirk> it can't be too hard
(18:15:16) <zgold> Griswold: Thats a valid question
(18:15:27) <zgold> Griswold: Maybe we convert to %s based on questions?
(18:15:30) <Craig73> winspool: a base set of questions will be very common / for 90% of the apps.  But large apps like autodesk need specific questions.  So the exceptions shouldn't outnumber the common questions too much
(18:15:35) <zgold> Aka an app works 70% based on this detailed criteria
(18:15:37) <Griswold> zgold, Could
(18:15:56) <Griswold> Hmm, I assumed you meant %s as in "convert to string"
(18:16:01) <Griswold> ie. printf()
(18:16:03) <zgold> haha
(18:16:09) Ingram_AV (n=Nathalie@145.94.201.52) left #winehackers
(18:16:12) <zgold> I meant like offering two scores
(18:16:20) <zgold> Installation score X/100 and Runtime score: X/100
(18:16:20) <Craig73> zgold: I still like the coloured ratings, just augment them with a percentage
(18:16:23) <Griswold> Yeah, at the very least an install and running score
(18:16:35) <zgold> Craig73: maybe we could tier the colors
(18:16:37) <Griswold> Could calculate the color by the percentage
(18:16:38) <austin987> agreed, but I'd say keep colors
(18:16:43) <Griswold> Ie. not have discrete colors, but more a scale
(18:16:44) <zgold> 100$ = platinum, >80$ = gold etc.
(18:16:50) <zgold> 100%, 80%*
(18:16:55) <Griswold> So you could have a "more gold" app
(18:16:56) <Craig73> zgold: agreed.... colours are easy / percentages for the nerds ;-)
(18:17:04) <zgold> Craig73: I'm OK with that.
(18:17:15) <zgold> So we have one idea: Use a scoring scale and base colors off of scores
(18:17:17) <zgold> More ideas.
(18:17:29) <vpovirk> for most apps we will probably have only one report
(18:17:32) <Griswold> Maybe make a bar?
(18:17:34) <vpovirk> for a given version
(18:17:44) <Craig73> I'm assuming that all questions will be equally rated?
(18:17:54) <Griswold> And have the install color "rendered" first
(18:18:00) <zgold> Craig73: i doubt that
(18:18:12) <vpovirk> so 100% can easily mean "one guy who tried it a year ago"
(18:18:16) <Griswold> So if it is 70% install, 40% run, it might look like:   xxxx===----
(18:18:19) <Griswold> Cept with colors
(18:18:33) <zgold> Griswold: perhaps
(18:18:41) <zgold> Alright so
(18:18:43) <zgold> Lets look at it this way
(18:18:46) <zgold> If you're a user
(18:18:46) <Craig73> I think older test reports should have a lower weight
(18:18:55) <zgold> who knows nothing about their pc, they likely couldn't even get into this IRC Channel
(18:19:02) <zgold> you want to run a windows game
(18:19:13) <zgold> your geek friend tells you, This game is rated "#####" by Wine
(18:19:19) <zgold> what information should that "####" convey
(18:19:29) <Craig73> That it runs
(18:19:41) <Craig73> so what if it installs... running is ultimately what is important
(18:19:43) <vpovirk> can I make it run on my computer?
(18:19:44) <winspool> Craig73: That's why the maintainer schouls select most impotant features. To get a base, what is more important, and what not.  => calculating rate
(18:20:06) <vpovirk> it's impossible for appdb to answer that question, but that's what I want from it
(18:20:10) <Craig73> (winspool: OK)
(18:20:27) <Griswold> vpovirk, Also how easily you can make it run..., not just is it possible
(18:20:31) <YokoZar> Why even have a rating for a test result?  Why not that be solely a maintainer thing to pick?
(18:20:35) <vpovirk> no, I just want to know if I can make it run
(18:20:42) <vpovirk> the answer is specific to me, personally
(18:20:49) <vpovirk> maybe I don't have a copy of windows to run the installer, maybe I do
(18:21:11) <Craig73> yokoZar: the test result is not the rating, it's all of the ratings combined by formulae
(18:21:26) <Craig73> yokozar: to get rid of the inconsistency of ratings
(18:21:29) <vpovirk> or maybe I don't have the technical skills to build wine with a patch
(18:21:40) <zgold> So i think "###" needs to convey that me, some silly user, can make the app work, or that i'll need to sit down and buy a pair of glasses and get friendly with the terminal and call my geekf riend often
(18:21:41) <YokoZar> Craig73: But applications themselves will have different formulas with different test cases, right?
(18:22:01) <zgold> Maybe a difficulty scale?  Combined with a total workability?
(18:22:16) <zgold> Aka. this app is rated "Working" but "Hard"
(18:22:33) <Subdino> likewise for dlls overrides: if I don't have a licenced version of windows I can't legaly get some needed dlls. that's why I suggested some "2d" scores, one axis being how "pure" the environment is, the other being the score
(18:22:47) <vpovirk> ooh, "Not Working", "Mostly Not Working", "Mostly Working", "Working"
(18:22:47) <Craig73> YokoZar: no - it should be a general formula, add up all the pass and failures, perhaps a multiplier for key features, perhaps a divisor for old test reports, ignore not tested (except for highly untested apps)... pretty generic formula
(18:22:58) <winspool> We forgot the question, if that app/game is a full version or a trial/demo (with DL URL)
(18:23:09) <vpovirk> that's a version thing
(18:23:20) <zgold> winspool: legit
(18:23:29) <winspool> vpovirk: perfect / good / partial / failed
(18:23:40) <vpovirk> partial?
(18:23:49) <Craig73> if it's difficult to install, it shouldn't get a platinum rating
(18:24:04) <vpovirk> if someone just says partial, I think of that guy in the office with the weird keyboard
(18:24:16) <winspool> vpovirk: perfect / good / lots of things does not work / failed
(18:24:16) <Griswold> Unless we have two separate ratings
(18:24:22) <Craig73> if it requires a tonne of work arounds / it shouldn't get gold
(18:24:40) <vpovirk> good has sort of the same issue; it's not clear to me that it's telling me how well something runs
(18:24:53) <vpovirk> with "Working" I know exactly what it means without thinking
(18:24:54) <YokoZar> Where will this rating be displayed?  I'm shy of putting anything automatically calculated in main view when an app has a maintainer who thinks different
(18:25:53) <Griswold> YokoZar, Yeah, but "Gold" doesn't mean the same to every maintainer
(18:25:57) <YokoZar> Since there will be different test reports saying different things
(18:25:57) <Griswold> Which is part of the problem
(18:26:08) <Craig73> YokoZar: why does he think differently?  are there test cases missing or misweighted?
(18:26:16) <YokoZar> Griswold: Of course, that's a huge problem.
(18:26:30) <YokoZar> Craig73: well suppose 2 testers tested it, and it worked for one and not for the other.
(18:26:47) <Griswold> If we keep the ratings, we need to defined what they mean
(18:26:53) <winspool> vpovirk: works perfect (without issues) / most things works / lots of things does not work / not usable
(18:27:03) <YokoZar> Griswold: we could put a textual explanation right into the interface that both the maintainer and the user will see
(18:27:12) <jeffz> Griswold: they are defined already
(18:27:13) <vpovirk> winspool: I like that better
(18:27:18) <YokoZar> Griswold: and we could clarify (eg "Wine version 1.1.14 with no special modifications")
(18:27:19) <Craig73> YokoZar: he should have a way to disregard test reports regardless, that way he can purge bad reports
(18:27:29) <vpovirk> whether it ran, and whether the features worked, are two different things
(18:27:30) <Griswold> jeffz, If an app doesn't install is it garbage?
(18:27:35) <YokoZar> Griswold: I don't mean in the AppDB ratings page, I mean right where it says Gold on the application page
(18:27:44) <jeffz> Griswold: yes, if you can't install it with Wine, how are you meant to use it?
(18:27:54) <Griswold> But then someone uses "winetricks dotnet" and it works perfectly?
(18:28:10) <jeffz> Griswold: that means it installs.
(18:28:28) <Craig73> griswold: souns like it :-)
(18:28:30) <vpovirk> so if the vast majority of people can't get it to run, or can't get any features to work, it's "Not Working"
(18:28:46) <Griswold> How about something like changing UseGLSL?
(18:28:52) <Subdino> jeffz: the installer might require dotnet
(18:28:56) <vpovirk> otherwise, we can only look at the people who did get it to run when deciding how well it works
(18:29:06) <Griswold> That allows an app to work that wouldn't otherwise, and it works really well afterwords?
(18:29:16) <vpovirk> the other part plays into the "Hard" bit
(18:29:22) <jeffz> Subdino: that means it would be Gold, since you've used dll overrides or third party software
(18:29:50) <jeffz> Subdino: or greater than Garbage anyway, if that succeeded in making it work
(18:29:51) <Craig73> jeffz: I'm not sure why not - WINE is not implementing all third party stuff
(18:30:00) <Griswold> The apps that are difficult to install or get running, but once running run well, what are they? Gold?
(18:30:15) <vpovirk> I'm sort of hoping we have two ratings..
(18:30:20) <Craig73> jeffz: I think the DLL overrides are what impact the status... because that is what WINE should be doing
(18:30:20) <Griswold> Same
(18:30:39) <vpovirk> that would be <zgold> Aka. this app is rated "Working" but "Hard"
(18:30:39) <Subdino> jeffz: I would go further in my idea, and rating it at the same time "garbage" in "pure" mode, and "bronze/gold" in "tainted" mode, or whatever it's named
(18:30:54) <jeffz> Craig73: sure, that's why it's not platinum, but it means that it works flawlessly if you use 3rd party stuff.
(18:31:24) <Griswold> Also, what about FPS issues for games?
(18:31:24) <winspool> Griswold: UseGLSL is something, a maintainer should select for the app to enable the question.
(18:31:25) <Craig73> jeffz: except WINE will never implement DotNet (at this point) so why drop the rating?
(18:31:42) <vpovirk> well, you have to assume that some people will send in reports that they did nothing special and it failed, and some people will send in reports that they did something special and it worked
(18:31:48) <Subdino> I would imagine at least 3 levels of "purity": pure, requiring redistributable files, requiring some licence
(18:31:49) <vpovirk> that's perfectly valid
(18:31:50) <Craig73> jeffz: if anything - WINE should make it easier to install legit third party apps.
(18:32:05) <Griswold> For some, a drop from 60 fps to 40 fps is horrible, for others (the more casual gamers), they will be just as happy
(18:32:20) <jeffz> Craig73: I don't think anyone is trying to make it harder to install third party apps in Wine.
(18:32:22) <YokoZar> It sounds like it would be nice if there were some way of relating both fiddling required and how well it works
(18:32:26) <Craig73> subdino:  but these apps wouldn't work on Windows without dotNET ... so why are they not platinum
(18:32:32) <Craig73> why is it not "pure"
(18:32:38) <Craig73> (users don't care)
(18:32:54) <Subdino> right
(18:33:27) <vpovirk> windows doesn't require you to install .NET
(18:33:40) <Griswold> Is anyone here in favor of keeping the current single-rating system w/ garbage/bronze/silver/gold/plat.?
(18:33:43) <Craig73> Griswold:  perhaps we need a rating for how fast the app runs  (ie - iTunes works but is horribly slow on my machine)
(18:33:54) <Craig73> I am
(18:33:58) <KittyCat> should appdb then be a place to tell users how well the game runs, or a place to log how good wine is behaving with various apps?
(18:34:02) <Craig73> griswold: I am
(18:34:03) <vpovirk> I would think features should cover fps issues
(18:34:05) <YokoZar> Griswold: yes I am, but we should have it be maintainer only not tester
(18:34:17) <KittyCat> ie. some problems may be indicative of a horrible problem in wine, but which may not impact user experience much at all
(18:34:17) <vpovirk> people who care would say they "tested" the fps, people who don't care would ignore it
(18:34:19) <YokoZar> Griswold: if there is no maintainer we just tell the user he can look at the test reports and become a maintainer himself
(18:34:31) <Griswold> vpovirk, Yeah, maybe speed should be in features for apps that it pertains to
(18:34:35) <zgold> Right, back
(18:35:01) <zgold> So, lets move on
(18:35:11) <Craig73> (default in survey should be "not tested")
(18:35:12) <Griswold> Should we do anything about speed issues?
(18:35:14) <zgold> Did we come up with any alternative ideas for displaying ratings?
(18:35:22) <YokoZar> zgold: yes yes yes
(18:35:29) <YokoZar> Can I present a doc file to you all
(18:35:30) <zgold> YokoZar: can they be precisely stated?
(18:35:53) <YokoZar> http://tuzakey.com/~scott/appDB-mockup.odt
(18:36:06) <Griswold> Most seem to want to keep current one-rating system, where as I would like 2 separate ratings for install and run  (unless I misunderstood)
(18:36:16) <Griswold> But the idea is that only maintainer can set the rating
(18:36:29) <vpovirk> only one person here wanted one rating?
(18:36:33) <zgold> Griswold: I like the one rating, but two scores?
(18:36:41) <zgold> I mean as far as a user is concerned
(18:36:43) <Griswold> Ok, sounds good
(18:36:46) <zgold> an app which installs but doesn't run is useless
(18:36:50) <Craig73> Griswold:  perhaps on the main page - it shows one rating.  Then when you go into the app, you see the rating broken down into install and run
(18:36:56) <Griswold> How about pick the lowest of the two?
(18:37:07) <winspool> vpovirk: It's required to install .Net on Windows. Most systems have it already in the meantime. (.Net in XP is an extra on the CD)
(18:37:12) <zgold> Griswold: I think we should have two ratings, install and run, and another (weighted) "overall"
(18:37:26) <zgold> And the overall would reflect the fact that if the app doesn't install its lowered etc.
(18:37:39) <Griswold> Hmm, that sounds ok
(18:37:40) <Craig73> zgold: agreed
(18:37:49) <vpovirk> depends whether you think of .net as a windows component that we are missing, or simply a thing you install
(18:37:53) <YokoZar> What about just having the main rating apply to Wine without special modifications, then have the other information conveyed elsewhere
(18:37:57) <Craig73> zgold: there was a great mapping to user language up above as well...
(18:37:57) <vpovirk> same could have been said of IE before we got the IE replacement
(18:39:07) <zgold> Can you find that Craig73 ?
(18:39:22) <Griswold> I think YokoZar is onto something about a user being able to just click a bar and give data that way
(18:39:22) <Craig73> yes... give me a sec  (I think winspool said it)
(18:39:35) <Griswold> Wonder if we can somehow merge that with the two sections of Q that we had before
(18:39:43) <Griswold> Maybe allow it, but give it little weight?
(18:39:56) <Craig73> "(07:26:51 PM) winspool: vpovirk: works perfect (without issues) / most things works / lots of things does not work / not usable"
(18:40:04) <vpovirk> is "rating" different from "report" here?
(18:40:21) Thunderbird quit ("Leaving")
(18:40:24) <YokoZar> Griswold: yeah, the idea is we'd get lots and lots of these reports for at a glance figuring out those 2 main problems a user had earlier ("will it work" and "what am I in for")
(18:40:26) <vpovirk> i.e. I can say it requires an annoying amount of tweaking without taking a full survey?
(18:40:35) <YokoZar> vpovirk: yeah
(18:40:42) <YokoZar> vpovirk: if you want to do a full test you click the test tab
(18:40:49) <YokoZar> vpovirk: and then do all the survey we designed earlier
(18:40:59) <Griswold> If you do a full test, you should be rewarded by having your voice count for more
(18:41:00) <vpovirk> I can't decide if I like that or hate it
(18:41:02) <YokoZar> vpovirk: the moderator can then look at that and update the notes/howto and such on the main page
(18:41:33) <Craig73> I like the two ratings on the apps page / just not for the front page where the apps are listed
(18:41:54) <zgold> So, what about the issue of what to do with current data?
(18:41:55) <Griswold> Front page I am less concerned about
(18:42:10) <YokoZar> zgold: well, current data will expire anyway as Wine changes.
(18:42:11) <Griswold> Keep current data
(18:42:15) <Craig73> zgold: keep the rating, just degrade it with time, older wine version, or newer tests
(18:42:16) <Griswold> But just give it low weight
(18:42:23) <YokoZar> By "expire" I mean lose relevance
(18:42:49) <vpovirk> that will most likely give us a better summary than attempts to generate from tests..
(18:43:01) <vpovirk> or relying on moderators
(18:43:14) <Craig73> zgold: and the appmaintainer should be able to disregard the test completely... as with all new tests, should they be suspect (rare cases hopefully)
(18:43:34) <zgold> Alright... we've been at this for almost 2 hours
(18:43:38) <vpovirk> it worries me that there's no way for the resulting summary to say "It just won't work"
(18:43:59) <zgold> People have stamina to keep going or we'll do this again next week?
(18:44:12) <Craig73> vpovirk: my only concern is I have seen appdb entries where people give a gold rating to something they only installed and didn't test much if anything
(18:44:35) <zgold> Craig73: that would come out in the new per-feature ratings
(18:44:39) <YokoZar> vpovirk: well they can click the bottom of the bar in my mockup ;)
(18:44:53) <Griswold> zgold, I can keep going
(18:45:02) <vpovirk> oh wait, it is there, heh
(18:45:08) <YokoZar> zgold: I can keep going, but I think we may have lost people that will be here next week.
(18:45:29) <zgold> I mean, we have enough data to begin some coding i think
(18:45:30) <vpovirk> Craig73: I think the hope is that if we get enough people doing ratings, we can just take the median and they won't matter
(18:45:31) <Craig73> zgold: one key thought on ratings... could we somehow split out the games ratings on the main page.  When people see the "top10" platinum apps are all games, we kind of loose credibility to some reporters
(18:45:42) <Craig73> vpovirk: OK
(18:45:49) <zgold> Craig73: Thats another discussion I think
(18:45:56) <zgold> All thats really missing is who is going to do all this?
(18:45:58) <Griswold> Also, do we want to allow testers to submit info about current Git Wine?
(18:46:01) <zgold> I'm up for it, but it will take a month ro two
(18:46:05) <jeffz> how about as an alternative for people who don't want to answer many questions, just a general question, "overall, were you satisfied or dissatisfied?"
(18:46:07) <Craig73> zgold: agreed... but it's a key output of the rating system
(18:46:11) <Griswold> Ie. maybe a fix *just* went into Git to fix a game
(18:46:19) <YokoZar> jeffz: see my mockup (http://tuzakey.com/~scott/appDB-mockup.odt)
(18:46:37) <Craig73> I'd like to see a couple of mockups before we code... just to see how we feel
(18:46:37) <Griswold> Or should we make them wait till the new version is released?  (or just use incorrect outdated version)
(18:46:41) <zgold> Craig73: I'de argue individual reports are more important, but doesnt matter
(18:46:58) <jeffz> YokoZar: ah, that looks simple
(18:46:59) <YokoZar> Yeah I think we should discuss this again next week...give us time to make some mockups and update the wiki and get more feedback
(18:47:20) <Craig73> zgold: yes
(18:47:22) <Griswold> Should probably do it at a different time next week so others have a chance to join in
(18:47:44) <YokoZar> Is earlier or later better?
(18:47:48) sedwards (n=sedwards@reactos/board/sedwards) joined #winehackers
(18:47:48) ChanServ sets mode: +o sedwards
(18:47:50) <Griswold> No idea
(18:47:51) <zgold> So should anything take place between now and then?
(18:47:54) <YokoZar> If we do it earlier we might get more codeweavers people
(18:47:55) <Craig73> isn't next weekend valentines day?
(18:48:02) <Griswold> I am thinking we can code some stuff up
(18:48:05) <YokoZar> Craig73: next Saturday
(18:48:09) <sedwards> evening
(18:48:15) <zgold> Griswold: you  up for coding this w/me?
(18:48:22) <Griswold> zgold, Either way
(18:48:32) <vpovirk> as far as people in the codeweavers office, I don't think you'll get more people who care
(18:48:33) <YokoZar> sedwards: of course you want evening ;).  I slept till 3pm today, woke up and came to this meeting ;)
(18:48:36) <Griswold> I am working on some cursor patches atm, but I can take a break from that
(18:48:48) c_korn quit ("Verlassend")
(18:49:14) <Craig73> Are we doing our own mockups as show and tell ?  :-)
(18:49:16) <sedwards> hehe no I mean good evening
(18:49:20) <sedwards> what are y;all working on?
(18:49:25) <YokoZar> AppDB
(18:49:25) <zgold> Griswold: I think we should do mockups first
(18:49:29) <winspool> The mokup from YokoZar looks nice, but with only the one-click rating, we will get less test reports.
(18:49:31) <Griswold> zgold, ok :)
(18:49:32) <sedwards> thats cool
(18:49:34) <zgold> Griswold: Some sample surveys and some sample ratings / outputs
(18:49:41) <Griswold> Yeah
(18:49:59) <Craig73> OK, thanks guys / later
(18:50:05) <YokoZar> winspool: We can encourage detailed test reports by making the test report page a bit more interesting.  The key here is we're losing a lot of data from people who aren't doing tests :)
(18:50:08) <Griswold> sedwards, Working on fixing up AppDB :)
(18:50:11) <zgold> I have to run, will take more notes and update wiki later
(18:50:13) Subdino quit ("Zzz")
(18:50:15) <Griswold> Throwing ideas 'round
(18:50:20) <YokoZar> All right.  Cheers everyone.
(18:50:25) <vpovirk> YokoZar: can I trouble you to poke me when the next meeting starts? I'm pretty sure I'll forget it again
(18:50:50) <YokoZar> vpovirk: Sure.  Are we eFriends on facebook?
(18:50:59) <winspool> And one or two test results must always be visible. Most users wont switch the tab, when they can click and go out.
(18:51:05) <vpovirk> no, I don't have a facebook account
(18:51:06) <Griswold> Meeting over?
(18:51:09) <Craig73> YokoZar / WinSpool : I was hoping that as new test cases are added, when someone tests they are presented with the additional questions.  Thus the quality grows over time.  The app maintainer can add a bunch too.
(18:51:14) <vpovirk> I'm either vpovirk or madewokherd on this channel
(18:51:23) <YokoZar> roger roger
(18:51:39) <YokoZar> So same bat time, same bat place, or diff time?
(18:51:56) <Griswold> zgold, Same time next week?
(18:52:00) <vpovirk> I don't think we'll find out today :(
(18:52:13) winspool is now known as winspool_zZz
(18:52:25) <winspool_zZz> by by ...
(18:52:28) <jeffz> bye
(18:52:29) <Griswold> Guess we can just bring it up on wine-devel
(18:52:35) <Griswold> When the next meeting should be
(18:52:56) <Griswold> YokoZar, Something unrelated:  How hard would it be to make a GCC 4.4.0 deb for Ubuntu?