Log of the meeting 18th feb 2008

* phomes hopes for a few more to show up in the last minute...
<-- triton has quit (Read error: 104 (Connection reset by peer))
--> andreasr (~andrearo@bacchus.pvv.ntnu.no) has joined #gnome-games
<phomes> yay :)
<andreasr> Hello ;)
 Nice to see you here. Is this the right time?
<madewokherd> I think so
<phomes> yes
<jordimas> Hello folks. I'm sick today. I may not be very active (:-
<phomes> okay. The agenda is pretty long so we better get started
--> triton (~triton@acb1-84-91-17-39.netvisao.pt) has joined #gnome-games
<phomes> the feedback on my post is sort of spread out on some of the other points so lets skip that one for now and start with the highscore
<andreasr> sure
<jordimas> ok with me too
<phomes> I'm all for the usage of ggz for this. But we need to talk about how to do this technically
<andreasr> Is joref here?
<andreasr> He should probably be able to say someting on behalf of GGZ?
<josef|samba> you mean josef? that'd be me :)
<andreasr> Yes, sorry ;)
<josef|samba> ok, some bits about highscores
<josef|samba> right now GGZ handles simple wins/loss/ties systems (including forfeits), and also highscores and player ratings (e.g. Elo in Chess)
<josef|samba> if other systems are needed, they would need to be coded first
<andreasr> does it handle both single-player and multiplayer games?  (you say "wins/loss/ties")
<josef|samba> GGZ only handles multiplayer, either with or without teams, although one of the players could be a bot
<josef|samba> it does however require the game server to report the scores, unless you use something along the lines of ggzcollector (which is still in playground)
<andreasr> ok. But could we somehow use GGZ to store highscores for all the games (both single-player and multiplayer)
<andreasr> hm...
 I liked the idea of using ggzcollector to submit highscores
 to some central highscore server
<phomes> yes
<josef|samba> it could be brought up to production use as part of ggz-python
<andreasr> sure
<triton> what about security?
<phomes> it would be nice to have different highscores for each release. Difficulty might change from release to release
<andreasr> true.
<josef|samba> basically, there is none when talking about absolute security... even with server-side score submission, people can still have scripted agents playing for them without anybody noticing
<andreasr> The KSirtet system seems to work:  http://ksirtet.sourceforge.net/stats.php?view=scores
 ikke tid
 Ops, wrong window.
<madewokherd> sorry I can't talk about technical issues, are gnome-games high scores really this exciting that it's worth having centralized high scores?
<andreasr> yes!  :)
<phomes> I think so too
<andreasr> But if it too difficult to implement, then maybe it's not worth it...
<phomes> well. Global highscore I don't care too much about. But I want to be able to compete with my friends
<andreasr> good point.
<phomes> high scores of my IM-contacts is what I want to see
<andreasr> josef: ggzcollector inserts data into the matches table in a GGZ database.
<andreasr> Do you have any examples of where this data is used?
<jordimas> Competition is a key issue in the success of many games. Like it or not.
<andreasr> (for highscores)
 where highscores is displayed?
<josef|samba> andreasr: not really, the tool had only one commit in 2005 so far, I didn't yet think of whether and how to separate such scores from server-side submitted ones
 andreasr: but we do have examples where the 'matches' table gets used in general, of course
<andreasr> ok
<phomes> lets aim for ggzcollector for high scores then :)
<andreasr> That sounds like a good plan.
<josef|samba> any python hackers here? we could do with some reinforcements in such a case
<andreasr> yes
<phomes> I can do some python as well
<andreasr> Perhaps I could try to submit some patches to ggzcollector
<josef|samba> good
<phomes> we can talk more about that in #ggz later. We better get on with the agenda
<josef|samba> ok
<andreasr> ok
<phomes> next up: bugs to fix before 2.22
<andreasr> Slow gnometris theme :(
<phomes> the theme in gnometris needs to be fixed
<andreasr> where was that bugreport again?
<phomes> https://bugzilla.gnome.org/show_bug.cgi?id=466939
<andreasr> We could either rollback to the old-and-ugly theme, or fix the performance problems...
 Comment #21 sais that "The lag problem is because we are saturating the X server with Xrender ops."
<phomes> there is a lot of gradients and that makes cairo cry
 from the talk carl worth gave at lca this will be fixed in cairo soon
<andreasr> Do this can be blamed upon carl worth?
 :)
<phomes> maybe not... :) but the theme will probably be smooth with later versions of cairo
<andreasr> Perhaps some of the gradients could be removed? eg. background, and some of the block-grandients?
 That would probably speed things up a bit
<phomes> but for now we should either revert or maybe see how a raster version would look?
<andreasr> revert is the quickest way to solve it.
<phomes> we could try that. My artistic skills are very low so someone else would have to look at that
 yeah. I think going back is best for 2.22 and then lets see how things are with future versions of cairo
 do we agree to go back?
<andreasr> fine with me
 next point on the list then? It's 00:30 here now, and I've got to sleep soon...
<phomes> I assume that everyone else agrees then
 we have some new crashers for sudoku but I'll try to look at them
 next: PPA's
<andreasr> PPA?
<phomes> ubuntu packages that we set up our selves. I'll get a link
 https://launchpad.net/ubuntu/+ppas
 we can set one up to have something to give to our bugfix-hungry users
<andreasr> ok
 Are you able to do this?
<phomes> yes
<andreasr> Then that sounds fine with me :)
<phomes> there are nice documentation for it
<andreasr> URL?
<phomes> I don't have it handy
 it's also part of their developer week programme
<oojah> You should be able to get help freenode/#launchpad I'd imagine.
<phomes> I just wanted to hear if you thought it was a good idea
<andreasr> yes, good idea
<phomes> okay. We need to speed up a bit unless this is going to be a very long meeting
<andreasr> I can't see any problems with it, can you?
--> shana_ (~shana@80.172.48.6) has joined #gnome-games
<andreasr> It doesn't hurt to try. Maybe people will use it, maybe not?
<phomes> okay - lets do it and see how that goes
--> mauricio (~mauricio@bd21fc07.virtua.com.br) has joined #gnome-games
<phomes> next:  defining gnome-games
 kind of blurry one :)
<andreasr> not an easy task...
<phomes> I just want to set some limits
--- mauricio is now known as piacentini
<phomes> some games are getting request for advanced features
<madewokherd> I want minesweeper with a spaces cleared count instead of a flag count
 and wraparound
<andreasr> madewokherd: bugreport?
<phomes> we can make e.g. chess the best and most advance chess game ever. But do we want that?
<madewokherd> closed invalid iirc
<josef|samba> oi piacentini 
<madewokherd> or maybe wontfix
 I was just providing an example
<phomes> for the desktop-included games I think we should keep things simple
<andreasr> yes
<madewokherd> that was the reasoning
<piacentini> sorry for joining in the middle. Mauricio Piacentini from kdegames, will listen and maybe contribute in the ggz-highscores/shared data discussion
<madewokherd> it seemed reasonable
<phomes> if people want the fancy advanced stuff there will be other games that better fit their needs
<andreasr> piacentini - great, we talked to Josef about setting up ggzcollector to collect highscres information
<piacentini> andreasr: cool, I will get a log afterwards, hope we can collaborate on this
<josef|samba> is shared data already on the agenda or should it be scheduled for the end?
<phomes> shared highscore with kde would be very nice. Getting the difficulty to be the same might be hard though
<andreasr> The highscores for individual games should be different to each game, and separate from gnome-games and KDE. But the development and hosting of the system could be the same.
<phomes> if we could get it right I see no reason not to share high scores with kde
<andreasr> sure
<phomes> we'll see how that goes. Probably not the first issues to look at
<andreasr> but highscores for Aisleriot should be separated from kpat
<oojah> I see no reason not to share high scores for some games. Something like kreversi/iagno would be idea.
<andreasr> (or whatever they call the solitaire game)
 the reason not to share, is that the rules are not the same.
<triton> because the gameplay is different
 and not fair :)
<andreasr> Onlt if the rules are identical, could the highscores be shared.
<oojah> That's why I only said for some games.
<triton> at least have them separate, then you can mix them if you want
<oojah> Indeed.
<piacentini> IMO the collaboration itself is the most important bit, it sends a nice message and pool resources. I would keep game scores separate for now, to simplify
<andreasr> yes
<phomes> ok
 let's look at that somewhere down the road then
 this "defining gnome-games" got kind of derailed. Are there input on that?
<andreasr> If you get any more ideas, put them here: https://bugzilla.gnome.org/show_bug.cgi?id=509690
 What is the current definition?
<andreasr> "GNOME Games is a collection of sixteen small "five-minute" games in a variety of styles and genres for the GNOME desktop. "?
<phomes> none...
 used to be that
<andreasr> what do you think it should be?
<phomes> there used to be a policy to not use saving etc
 I think the most important think is to have a few entertaining "5 minute" games that are really well integrated with gnome
<josef|samba> There were some requests from jordimas and others to include more games - what about these? Or should the games always be the same?
<phomes> I'd like to have fewer games actually
<madewokherd> what we need is to get Jonathan Blandford back and have him decide all this
<phomes> I'd like to cut down on the number of games and focus on the quality of those
<shana_> how many games are there at the moment?
<phomes> 17
<shana_> that's... a bunch
<phomes> something more like 7 sounds reasonable to me
<madewokherd> most of the games we'd get rid of are not the same ones that are hardest to maintain, are they?
<phomes> no. But some are really ugly. And we clutter the menu
 we hold stuff back because we don't want the games-pack to grow in size
<shana_> if you're gonna cut down on the number, get a nice round figure so you can line them up nicely on the menu :)
<oojah> This is going to be a tricky issue. I'm sure everyone has different ideas about which should be included.
<phomes> indeed
<madewokherd> ultimately I think we really just need someone to decide this
<phomes> and I don't feel particularly well being "the new guy" to just cut some away
<madewokherd> there's very little chance of consensus, and if you try to go by what the group thinks everyone will want his/her favorite games to stay
<phomes> setting criterias for tangoed graphics, network support etc. will help
<jordimas> I would be good to involve the users in the decision somehow
 like web voting 
<andreasr> I don't think that the number of games should be reduced that much.
<madewokherd> I don't see a problem with removing games from the package as long as they're still accessible
<andreasr> Removing one game during the next development cycle would be possible.
<phomes> I don't want to trow the games away completely. A gnome-games-extra-games could be set up
<madewokherd> so what you really want is to split up the package?
<phomes> yes
 pick 7-8 games and run with those
 we can still maintain the others. Just make a separate release with them
<andreasr> If you split gnome-games into several packages, you loose many of the advantages of being in one package: shared code (libgames-support), translations, graphics, sounds etc.
<andreasr> So I'm against splitting it.  :)
<madewokherd> why do we lose shared code?
<chpe> you could simply alter the default set of games compiled by default;  with the others still available via  --with-games= configure option
<phomes> there has been requests to make libgames-support external anyways
<madewokherd> changing the defaults doesn't really matter if the games will be packaged the same way
<phomes> installing the games for those who want them is not going to be difficult for them. I just want gnome-games to be a collection of the best and most integrated games we have to offer
<madewokherd> (by distros)
<triton> i agree with phomes 
<madewokherd> I don't see why anything can't be shared if it's split
<triton> just have an extras package
<chpe> you double the release work
<phomes> that is right
<chpe> for no gain over --with-games=
<andreasr> exactly.
<madewokherd> you'd get an external libgames-support
<triton> chpe: that is fine too
<phomes> problem is that no one cares to use it
 no one = no distro
<andreasr> Red Hat uses it, I think
<madewokherd> cygwin uses it
 er
<chpe> triton: why? will -extra-games really see as many releases as necessary, e.g. to get updated translations out? making a release does require much time
<madewokherd> cygwin-gnome?
<phomes> and they don't care about including new games either
<madewokherd> something like that
<triton> chpe: then I guess having it as a compile time option is fine
 distros can package it how they want anyway
<andreasr> yes
<chpe> triton: thing is that we already do have that configure option :)
<triton> oh nice, then it's just a matter of changing the defaults
<phomes> by that rationale it should be okay to increase the games size as well?
<chpe> yes, imho :)
<oojah> If the defaults are changed to include only 7 games then the likelihood is that the vast majority of people won't have any access to those games at all as they won't be packaged.
<chpe> triton: problem is that those that aren't in the default set will get much less testing that way. probably unstable releases should compile all games
<phomes> we should package them as -extra-games
<madewokherd> but then maybe the 7 games will get more testing ;)
<phomes> problem is that distros seem to be non caring about what games are included
 they could just have included chess and sudoku themselves
<andreasr> most of them did, already...
<phomes> per default?
<andreasr> I would have to check.
<madewokherd> can we add something to the GPL that forces distros to bow to our wishes?
<andreasr> :)
 no
<phomes> he he
<oojah> I've certainly noticed iagno, gnect and gnibbles on most distros I've looked at.
<phomes> but chess and sudoku?
<andreasr> I've got to go. It's time to sleep here in Norway. Bye!
<oojah> Er... maybe I misunderstood the context. I guess now that you're saying they aren't included by default.
<phomes> okay. Night Andreas
<-- andreasr has quit (Lost terminal)
<oojah> I thought you meant they could have *only* included chess and sudoku.
<phomes> well. They got included in gnome-games and by that way in distros
 my point was that distros don't care enough to cherrypick games
<oojah> I agree.
<phomes> if we give them 17 games then that is what they include (except when there are legal issues)
--> josef|sa1ba (~josef@dslb-088-073-232-084.pools.arcor-ip.net) has joined #gnome-games
<phomes> but we are not going to do anything for 2.22 anyway so lets move on if there are no more comments
 externalizing libgames-support
<madewokherd> are there people who want to use it in games?
<phomes> there were some request
 s
<madewokherd> are they here right now?
<phomes> but I think they expect some fancy things are not going to be there
<madewokherd> has anyone tried to explain what they would use from libgames-support ?
<phomes> network support and fancy graphics (not there)
<madewokherd> they'd get fancy playing card graphics
<oojah> There is an awful lot of duplicated network code between the networked games. It would certainly be worth investigated whether it could be made split out into a library.
<phomes> the current lib has some nice stuff with clocks, boards, high score etc. If any games outside of gnome-games has a need for those it would be nice
 if they expect 3d graphics... well
<madewokherd> but we don't actually know if they do?
<phomes> no
<madewokherd> can we ask?
<phomes> definitely. They can comment on the logs and we will take it from there. I will mention it on the blog
<-- josef|samba has quit (Ping timeout: 600 seconds)
<phomes> if there is a demand I think we should do it
<madewokherd> I agree
<phomes> next: network games
<madewokherd> I wanted to ask something maybe slightly related but I'll wait?
<phomes> no no. Just ask
<madewokherd> are the python games able to use libgames-support currently?
<phomes> sorry if I move on too fast
 no
<madewokherd> would having an external library make that easier/possible?
<phomes> there are no python bindings. It's planned though
 to be honest, I don't know
<oojah> It seems unlikely it'd make much difference - someone just needs to do it surely.
<phomes> yeah
 more comments on this?
* madewokherd shuts up for now
<phomes> okay :) network games then
 LAN support is missing, but is this something we consider important to look at?
<oojah> I'd say that it would probably be quite a lot of work.
<madewokherd> having never used the network support at all, I cannot say >_>
<phomes> auto detecting that the computer next to me is ready to play a game would be nice. But to still play the game via ggz servers
<oojah> Ah, that sounds more interesting.
<phomes> probably more work than it's worth
<oojah> zeroconf support or similar.
<phomes> yes
 I'd like to start a game easily with im-contacts and local computers
<madewokherd> do we have any LAN im protocols?
<oojah> Yeah.
<phomes> I meant that those are the people I would play against. The guys at the uni and my friends from the list in my IM app
<oojah> I think im contacts are much more important than lan.
<phomes> I agree
 it's not as easy as it could be to start a game with an im contact. I had to get on the phone with a friend of mine to guide her
<oojah> ggz does have a pigdin plugin kicking around, but I don't know the status of it. I'm not sure it's in a useful state and it's not been developed much.
<phomes> would be nice to look more into that
 having your online friends available directly in a menu in the game would be optimal imo
 like: menu -> network game -> online friends -> peter
 game pops up on peters computer asking if he wants to play
 that kind of integration
<oojah> Yeah, that'd be neat.
<phomes> but that's more "stuff that would be nice to look at for 2.24"
<josef|sa1ba> if peter doesn't have a game running yet, you'll need integration with one of those presence services that currently evolve around galago, telepathy etc.
<phomes> yes. That was what I was thinking
 jason would probably kill me for having telepathy mentioned in #gnome-games :)
 any more comments on that?
<oojah> Slightly different topic - I'd like to make sure that all games use common servers where possible. So for example all chess clients used on ggz use the same game server at the moment - that's great. ggz reversi and kreversi both use the same server but iagno uses its own - less great :)
<oojah> Again that obviously only works for the case where rules are the same.
<phomes> yes
 there is a lot of work gone into a iagno next-version
<oojah> I saw the bug report - it was the talk of changing the network protocol that made me check what the situation was.
<phomes> if possible we should use the same server
 there is a ton of things in that patch and no real description. I lose all my energy when I start reading the patch. But I'm gonna take a better look early in 2.23 to see what we can do
<oojah> Yeah, it's a monster bug.
<phomes> what are the games in kde 4 with ggz support?
<oojah> E...
 (Er...)
 I'm not sure.
 josef|sa1ba: ?
<phomes> ok, never mind then
<josef|sa1ba> http://play.kstuff.org/ <-- they're listed here
<oojah> Thanks
<phomes> we should just make sure to use the same servers for those
<josef|sa1ba> KSquares will be added in 4.1, and maybe KFourInLines as well
<phomes> ok
 next is glade/gtkbuilder
 I want to port some more games over to maemo so gtkbuilder will be a problem for that
 chpe: you don't want glade for aisleriot at all right?
<chpe> that's right. one less dependency to worry about :)
<phomes> he he
<chpe> maemo4 has gtkbuilder support, but I still want to continue to support maemo3
<phomes> okay
<chpe> and anyway aisleriot's code just got renovated. the other games can still use it :)
<phomes> using gtkbuilder for the newly ported games should be fine though, right?
<chpe> yeah
<phomes> that's the plan then
 next is the new default cards
 are there any comments on using the tango-with-black-black cards?
 I would like to see tango where it makes sense to do so
 okay... Theming in general
 having some guides on l.g.o about theming would be nice
 what files to change, formats etc
 the games really need some love there
 and sounds that don't give you a heart attack like the "aiiiii" in robots
 okay, talking to myself here :) Bugzilla
 I have no idea if we can split that up
 chpe: do you know if it is possible at all? If it is something I can change then sure
<chpe> I think by talking to bugmaster to get gnome-games-{general,whatever}-maint@gnome.bugs aliases, you can then assign these as default assignee on some component(s) while leaving the default QA as gnome-games-main@.
<chpe> that way one can watch for all bugs, or just the ones for that component
<phomes> okay, I will look at that tomorrow
<chpe> I don't think we need aliases for all games though, but I would be interested in watching aisleriot bugs, but I'm not at all interested in the sudoku crash bugflood :)
<madewokherd> I'm already doing that with gmail filters
<phomes> we don't want you to drown in those duplicates :) 
<-- josef|sa1ba has quit (bedtime)
<chpe> right, client side filtering would be an option too
<phomes> will you do that client side or should I check if it is possible to do it on the server?
<-- shana_ (~shana@80.172.48.6) has left #gnome-games
<chpe> server side would be preferred, since then I wouldn't even get the mails instead of killing them after downloading them :)
<madewokherd> I'd use it
<phomes> okay :)
<madewokherd> (not that I care if google downloads them)
<phomes> any general comments for me? I'm very much the "new guy" so do let me know if there is something you want me to do
<madewokherd> be on irc more
 I like irc
<chpe> if anyone knows scheme, I'd appreciate reviews of the aisleriot patches that touch scheme code :)
 otherwise, no wishes
<madewokherd> >_>
<chpe> also it's time for sleep, bye :)
<phomes> I can read scheme but I think reviewing it is beyond me
<madewokherd> should I assume that if no one asks?
 (that scheme reviews are wanted)
<chpe> yes
<madewokherd> k
<chpe> I just know enough scheme to put together the dealable? implementations, but I'm not at all sure they're completely correct, for example. so I appreciate anyone pointing out the errors of my ways :)
<-- chpe has quit (going to sleep...)
<oojah> Bed for me as well- night.
<phomes> any other comments? I'll write something up and blog about it tomorrow. We should schedule a new meeting to follow up again later too
 night
<madewokherd> why can't we just chat on irc like normal people?
<phomes> we can :)
* madewokherd looks for thet dealable? bug
<phomes> I was sort of hoping that more would attend the meeting. We are spread across many timezones but an official meeting didn't quite help
 https://bugzilla.gnome.org/show_bug.cgi?id=445955
<madewokherd> I think normal (not specific times) irc is just easier
 I also don't think chpe has much reason to be worried about his scheme
<phomes> yes. But if there was to be some sort of consensus about what games to keep / new games to investigate then that would have been nice
<madewokherd> more people would create less consensus about that
<phomes> I'll hang around on irc more though
 yes but I do think some should be asked anyway
<madewokherd> I'm pretty sure the more you try to involve people, the more they'll push you towards keeping all the games
<madewokherd> and possibly adding new ones
<phomes> I'm not adding even more games that's for sure
<madewokherd> of course, saying this is just part of my evil plan to eliminate all games that are not aisleriot