Announcement

Collapse
No announcement yet.

What Should Valve Do For Linux & Open-Source?

Collapse
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • #16
    What should Valve be doing for Linux and open-source software?

    "Write Games"

    That is all.

    Comment


    • #17
      Be visible within the community

      It might sound odd, but just being visible can reduce friction and potential problems.
      It's really frustrating when people run into the FOSS community, do something amazing, then disappear.
      Maybe send a guy to some of the larger FOSS conferences (LinuxCon, OSCON, etc.)

      It also reassuring to know that you're aware of the next couple of evolutionary iterations.
      Windows has traditionally been very slow to evolve (haven't used windows in some 10+ years, so ignore this if not applicable anymore), with long periods (years) of stagnation then a sudden, large jerk forward.
      Linux is much more iterative, with lots of smaller changes over much shorter time-scales, and may require a different mind-set.
      This is particularly apt for closed source software where we can't just recompile against the latest libraries.

      Picking a project out-the-air just as an example; Wayland.
      Knowing that you're aware of it (because you're visible in that community) is very reassuring to those who wonder how Valve's software will evolve over time.

      Open-sourcing as much as possible is also a good way of ensuring your project stays on top of future advancements.
      I realise that maintaining control over the platform is important, but also understand that the more open a "closed" project is the less of a burden it'll be to support everyone.
      It should be possible for your project to be completely agnostic on a lot of plumbing infrastructure.
      Eg. Display managers, as mentioned above, but also sound, distro packaging or any other plumbing piece that can be swapped out for something else.
      It should be perfectly doable for example, to get the project into a state where any distro maintainer can create a package for Steam, or any of it's games.
      Then the burden of packaging is not Valve's concern, but that of the individual distributions.

      And lastly, communicate.
      The last thing you want is to get into a situation like nVidia is now or xFree86-of-old where the only way people seem to be able to get things changed is to scream to high-heaven until they're hoarse in the throat and frustrated enough to usurp you in favour of a grass-roots alternative.

      Comment


      • #18
        Originally posted by Michael View Post
        That's somewhat of a misconception... I put Valve in contact with Nouveau [Mupuf] months ago, but that they don't really need/want money. Their developers are currently still at university (and don't want to quit) or are happily employed elsewhere, so first they magically need to find some qualified developers. It's also somewhat also the case for Radeon; I tried to get Marek to talk to Valve, but alas he wants to go to university for another year.
        Offer them graduate positions?
        *cough* Steambox *cough*

        Comment


        • #19
          Keep the steam-client as open and flexible as possible, use sane library loading not locking it to the ubuntu/debian style and pointless version requirements like we see in many half-closed projects.

          pq

          Comment


          • #20
            Valve, use your influence to make AMD improving their open source drivers to an extend, that they're usable for modern and demanding games.

            ... Besides having the games, this is last thing I miss

            Comment


            • #21
              What Valve should do:
              1. Transparent cross-platform build with Valve tools
                The dev shouldn't have to invest time or energy even thinking about if it is worth his while to support Mac/Linux. It should be no more than a button press to build for all three insofar as is possible.
              2. Open bug tracker so people can search for issues, raise them, provide feedback, and track them
                Witness the blank cliff-face that is ATI/AMDs propriety driver support (admittedly this may be better now, but I haven't heard of an official bug tracker, just the unofficial one)
              3. Put pressure on Nvidia to provide docs to the nouveau guys without NDA's
                Encouragement toward AMD on their sterling efforts so far, and outright congrats towards Intel, who seem to be the most enlightened and productive. Intel do not just work on their own drivers, but (AFAICT) work on the core infrastructure too.
              What Valve should not do:
              1. Build a package for every distro
                But by the same token, don't make it an immense pain to get running elsewhere because of bad choices. Distro Community feedback should be heeded, because they are the experts.
              2. Write drivers, infrastructure, compositors, operating system level stuff
                This is not their core mission, and getting dragged into this will reduce the focus from where it is needed: The client, the tools, and the games.

              Comment


              • #22
                What about VALVE helping out the community to attract new contributors which Linux and VALVE can benefit from (mid to long-term)?
                ...One of the main issues to date concerning the graphics drivers IMHO.

                How that 'could' be done:

                • Creating something similar to GSOC and paying students to work on certain FOSS subjects.
                • More importantly - Besides providing money, it would be a major benefit if they would pay someone to professionally
                  write some tutorials how to get into Linux GPU driver programming. This would be a completely new concept which
                  might attract people with good Linux and programming knowledge but no knowledge in the hardware/driver business.
                  It's all about 'lowering the threshold'.
                  Let's face it, you have to pay someone to work on such documentation. It won't happen otherwise.
                • Also providing some resources to create the webpage infrastructure and host this on their web servers
                  would be much appreciated.

                I'm sure this would generate a very positive attention from the media which in turn will reach and attract more new contributers.
                This would surely benefit the Linux environment and VALVE.

                Comment


                • #23
                  I don't see why Valve should contribute to FOSS drivers (i ll be more than happy if they do of course). Its not their business in a way.

                  Make it run smoothly is their only "obligation" i think.

                  Comment


                  • #24
                    - make their software work on Linux, many linux flavors
                    - maintain it, make sure it keeps working
                    - try to squash bugs in important drivers (graphics and audio) and infrastructure software that have an impact on the software from Valve
                    - add platform icons in steam so people can see if their games will work on Windows, Mac, Linux,...
                    - work with Canonical to optimize (Unity) and fix (PulseAudio) things
                    - Buy and burn the ST3C patents

                    Comment


                    • #25
                      Originally posted by smani View Post
                      Contribute to KDevelop / QtCreator to make the platform more attractive for developers used to Visual Studio. While I personally can't stand VS, in the research lab here it is basically the only reason most are using Windows.

                      While QtCreator is very solid, it's code analysis is still somewhat basic.
                      KDevelop on the other hand has some very nice features here and there, but it unfortunately still is, in some circumstances, too crashy or unreliable to really be recommendable for productive use.
                      I'm one of those Visual Studio fetishists .

                      To be honest: We've a very deep problem when it comes to developing things for linux. Nearly every IDE out there is focused on one thing. Either GTK or QT. Or no GUI at all.

                      Someone should just do a IDE project from scratch. It should support every 'popular' language (C# (mono)/C/C++/python and so on) and at least GTK and QT for GUI stuff. And by support a language I mean Visual Studio level. A support site like the msdn for every language, code highlighting and suggestions, add-in functionality, GUI porting (you've a GTK application and can port the GUI to QT...more or less) and so on.

                      my2cents.

                      Comment


                      • #26
                        Originally posted by M1kkko View Post
                        I fired them off an email about this earlier today, but let me say it here as well: There are noticeable performance issues when it comes to gaming on a stock Ubuntu system.

                        Here's a quick look at some of the results of the tests I've conducted earlier this year: A comparison of different desktop environments, when playing a game on wine 1.4.0:

                        [...]
                        This is being sorted right now. Ubuntu 12.10 already has the first fixes in (Unity) while others are in development and landing soon (Compiz). The 3D performance of Ubuntu 12.10 will be *much* better with Unity, and they want to try to backport some of the stuff to Ubuntu 12.04.

                        Comment


                        • #27
                          pressure nvidia to make their closed source drivers even better.

                          And, like the rest of the posts, make all 11 of the team members work more on linux than on games.

                          Comment


                          • #28
                            An important thing I am pretty sure they will do: Provide their Source engine licensees access to their Linux code.

                            Another thing they could do is provide some incentives for games on Steam to come to Linux, like taking a smaller cut for the first month if a game launches with Linux support or something like that.

                            Documentation, examples and tutorials and open sourcing of some code they have used to easing porting of Source would also be pretty cool. Notice this doesn't imply open sourcing the Source engine it self.

                            Working with the OpenGL guys and similar committees concerning various standards to help push things along on that front would also help.

                            Comment


                            • #29
                              Originally posted by Detructor View Post
                              I'm one of those Visual Studio fetishists .

                              To be honest: We've a very deep problem when it comes to developing things for linux. Nearly every IDE out there is focused on one thing. Either GTK or QT. Or no GUI at all.

                              Someone should just do a IDE project from scratch. It should support every 'popular' language (C# (mono)/C/C++/python and so on) and at least GTK and QT for GUI stuff. And by support a language I mean Visual Studio level. A support site like the msdn for every language, code highlighting and suggestions, add-in functionality, GUI porting (you've a GTK application and can port the GUI to QT...more or less) and so on.

                              my2cents.
                              No, please not Yet Another IDE Attempt, which will end up seeing all the basic functionality other projects already reimplemented and then being abandoned when hitting the same problems other IDEs have (writing good code analysis is a royal pain!)
                              While QtCreator is very Qt/C++ focused, KDevelop also supports Python and PHP, and is not Qt focused (well, it has syntax highlighting for the occasional Qt specific keyword/macro, but that's about it).

                              Comment


                              • #30
                                Originally posted by Kristian Joensen View Post
                                An important thing I am pretty sure they will do: Provide their Source engine licensees access to their Linux code.
                                That would be an important one. I'll raise you the previously mentioned, "Provide developers with an easy-to-use cross-platform/cross-compiling toolchain which can generate Linux/Mac/Windows binaries from a given source tree with minimal fuss." If that's not feasible (Cross-compiling mac/win probably can't be done from other OSes), then just try to make the process as painless as possible.

                                Originally posted by Kristian Joensen View Post
                                Another thing they could do is provide some incentives for games on Steam to come to Linux, like taking a smaller cut for the first month if a game launches with Linux support or something like that.
                                I like that idea. Any game that launches with full Steam Play (Win/Mac/Linux) support would get a smaller percentage cut to Valve (which Valve hopefully makes up through higher sales volume).

                                Also, I'd love to see the post-launch statistics from their periodic surveys on Windows/MacOS/Linux hardware/OS statistics. We might get some real-world numbers on the number of people/gamers running Linux.

                                Comment

                                Working...
                                X