TL;DR A fully functional replacement for SteamOS on Steam Machines
If a gamer buys a Steam Machine and wants to use it in full freedom, what
advice can we currently give them? Uninstall Steam and any other proprietary
bits in SteamOS, leaving them with a weirdo Debian system. Or format and
install an existing libre distro like Trisquel. This doesn't provide the
console-like UX (User eXperience) of SteamOS and I think we can do better.
The vision for GOLD is to provide that UX, in the form of a gaming-orientated
GNU/Linux distro that meets the GNU Free System Distribution Guidelines. One
possible approach could be to investigate Lakka to see what components would
need to be removed to meet the GNU FSDG:
http://www.lakka.tv/
My understanding of SteamOS is that it aims to work like a console, providing
a large, simple GUI that can be seen and used easily from the couch, with a
wireless controller (mouse and keyboard, or game controllers). Steam serves
as this GUI. There is a standard Debian desktop as a fallback, in a similar
way to how CTRL-ALT-F1 terminals exist as a fallback on most GNU/Linux
desktop environments. But a desktop environment designed for sitting in front
of a desktop or laptop doing work does not provide the same UX, so a Steam
Machine minus Steam loses something important, for which there needs to be a
libre replacement.
Another aspect of Steam that we would want to replicate is the store
functionality. It needs to be easy to browse and download/ install games from
an online source, easy to uninstall them again if the player runs out of
space, and easy to buy bundles, or donate to developers, or contribute to
libre game crowfunding, and so on. This functionality could be limited to
free culture games, to encourage free licensing of more games. This would
also help to provide funds to those who develop free culture games, so an
industry can grow up around them, and people can make a part-time or
full-time living working on them.
The exact policies on the licensing of game art (graphics, audio etc) would
be set by consensus of the volunteers that join and work on the project (if
and when that happens).
One option would be for the distro to only supply free culture games from its
main repos, and have a separate repo(s) for games with 100% free code, but
non-free art. While the end goal would to support the playing of free culture
games on free code game engines, in the short term, a 100% free code platform
providing a roughly equivalent replacement for Steam would be the first
priority.
GOLD could have a libre fork of something like Lutris or PlayOnLinux as the
games library. Or, an existing media library app like Kodi could be extended
or forked to do the same job. There could be a couple of default desktop
options, as with Trisquel, something like KDE or GNOME for computers less
than 5 years old, and maybe a tidied up version of something like
Enlightenment 17/ Moksha for older computers. OpenBox could also be there as
an option for more advanced users, comfortable with remembering commands and
package names. The gaming library could have an interface for setting up
network games too, and maybe a default list of libre gaming servers that gets
regularly updates pushed through packages management, to encourage the
emergence of player communities around libre games.
Voice chat is an important part of the social experience of gaming. There are
existing packages like Mumble clients that could be integrated. Support for
sending notifications (e.g. "now playing 0 A.D. come join me") through
existing apps, maybe with a plug-in infrastructure that can support GNU
Social, Diaspora, Friendica, Hubzilla, Pump.io, Tent, chat systems like IRC,
XMPP, RocketChat or MatterMost, whatever there is demand from users for. This
is a nice-to-have, but it is a part of the PlayStation/ XBox experience that
Steam tries to compete with, and it would be great for GOLD to have it.
The GNOME project has been assembling game-related resources on this wiki
page.
https://wiki.gnome.org/Design/Playground/Games
My first impression is that it's still focused on a PC gaming experience
(sitting in front of a computer gaming with a keyboard and/or mouse), rather
than the Steamesque console experience (sitting on a couch with a screen on
the other side of the room and gaming with wireless controllers). That said,
it looks like a good project, and the work they are doing could well be
useful in the development of GOLD.
The ideas behind this proposal were first put forward in this thread:
https://trisquel.info/en/forum/could-lutris-fork-be-used-create-libre-replacement-steamos
...which seems to have got a bit bogged down in side discussions, so that's
why I've tried to aggregate the ideas here. I accept that it's not a
universally accepted opinion, but I think we need to at least consider what
would be needed to support VR gaming on GOLD. There are already open source
projects underway to build large chunks of the VR stack
(http://www.osvr.org/), so I think there's potential for similar
collaborative projects to set up to develop libre game libraries/ engines for
VR. I'd prefer not to dwell on this topic here, at least for now, as it's
already been talked around in circles a few times in the thread linked above.
