> Le mardi 1 septembre 2015, 22:50:57 Markus Koschany a écrit : Soo, Id' like to wrap this up. *) I'd keep the current "technical" page mostly as-is; as this is quite usefull as development dashboard (I'll add the TODO games too). Screenshots & long description would be dead weight here. As it would be quite deceiptfull for user not tracking git version to see there that $game is supported; but to realise that this game/version isn't support by their outdated package; I'll add a link to a frozen-in-time & reproducible version at file:///usr/share/doc/game-data-packager/list.html the disclaimer at the top would be quite different: - This is an automaticaly generated list of games supported by the upcoming release. + $wording "dpkg-parsechangelog --show-field version" on "dpkg-parsechangelog --show-field date", + list, for upcoming release please see http://pkg-games.alioth.debian.org/game-data/ *) As separate "marketing" page will be built, in the visual likes of http://blends.debian.org/games/tasks/finest . Instead of listing every single game as in the "technical" page, games would be grouped by franchise of similare games: - doom, doom2, final-doom, compet_n together; but not with doom3 & doom3 bfg as they are something else - zork games together , explaining the evolution from text games to visual ones .... with one screenshot per game & list of engines used in the description. This needs some new "franchise:" hint in G-D-P .yaml files. (trivial) The description-per-franchise shared by several games would be stored somewhere-else. The long descriptions of one-of-a-kind games (Tyrian, ROTT) would go in the indivual .yaml; and added to the generated .deb's too. > > This feel like a big regression: > > I think the implementation details are up to you. I've got a bad gut feeling about this. > I could also imagine to use debconf, to ask the user a couple of questions > and drop root privileges where appropriate. The point was that a customized > metapackage could combine features of g-d-p with normal Debian packages, > that are already available in the archive. It is simply another high > level layer and > > apt install games-scummvm > > is probably easier to remember for users than a dozen g-d-p commands. *) I don't think it's really a problem if users must endure some minor incovenience to get non-free software running on their system; and I prefer having separate packages that does what is says on the tin like "geoip-database-contrib" that having a package called "games-scummvm" that unexpectedly starts downloading 1GB of demos/free'd full games (Dreamweb) when installed; even when user already have the full version of these games somewhere; but they had to specify the path by hand. If common RaspBian users are willing to go to these step-by-step instructions guides: http://www.instructables.com/id/PIrate-The-Secret-of-Raspberry-Island/?ALLSTEPS http://www.codingepiphany.com/2013/06/23/running-gog-com-scummvm-games-on-raspberry-pi/ they'd find awesome if they are told that they can buy the game on their Windows PC and run G-D-P & lgogdownloader in-host on their Raspberry with a single command & without having to mess around with locating files in C:\someting and moving those around with USB sticks & editing scummvm configuration files. > > apt-cache search <somegame> will also list game-data-packager ; > > maybe it's more a documentation problem than a technical one. > > Yes, it is both a documentation and usability problem. I think a simple > GUI would be helpful for g-d-p too. I was thinking of starting with something a bit lame, that'd look like a 10 years-old xcdroast with a big terminal widget. Then I found this...[1] [1] https://lists.debian.org/debian-devel-games/2013/05/msg00041.html I guess it's the way to go instead of holding it back forever. > The question is whether you want to promote the contrib games > aspect of Debian in general (free engines but non-free data) > or just a tool like g-d-p. I'd prefer users using Debian-provided scummvm than using some scummvm.exe in Wine; this way the engine continues to get regularly updated and they get a better overall exeperience. I want to promote the contrib games aspects and g-d-p is a mean to achieve that goal. > If it is just g-d-p, I would invest the time in a good man page, examples > and a GUI and forget about the rest. I'm not good at writing documentation; the man page can certainly be improved. > I think it is more interesting to write documentation, articles and > guides about the bigger picture. Yes... but I've read this [2], and I think the "Apparently we haven't found this consensus yet." is still valid; for me a free engine without free-data is useless in main & should go to contrib (that includes residualvm, maybe rbdoom3bfg...) [2] https://lists.debian.org/debian-devel-games/2014/02/msg00086.html If we can at least come up with a basic description of the situation that includes what is considered common knoweledge on d-d-games and everyone agrees on that'll be a good start ! Regards, Alexandre Detiste
Attachment:
signature.asc
Description: This is a digitally signed message part.