Announcement

Collapse
No announcement yet.

550 Days Later, UT3 Linux Appears Dead

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

  • #31
    Originally posted by Kano View Post
    OpenGL based games can run very fast with wine, even before the Doom 3 Linux binary was out wine was able to run it - and not even slow. But games like UT3 don't have go a OpenGL renderer anymore - if it would have got one, it could run MUCH faster than using the D3D renderer using wine. To port a game without OpenGL render path is certainly more demanding than when it is already there. Not sure if they needed that already for a PS3 port, if so they could add it to an update too...
    It's, more often than not, only slightly harder than the OpenGL port (which is often largely a matter of plugging in a new video player engine, a new sound subsystem (if they didn't use a supported one with a license or something like OpenAL), and going looking for the oopses that using VC++ will let creep into your code.)- it's not quite as hard as one would think it is.

    Comment


    • #32
      Originally posted by Svartalf View Post
      It's, more often than not, only slightly harder than the OpenGL port (which is often largely a matter of plugging in a new video player engine, a new sound subsystem (if they didn't use a supported one with a license or something like OpenAL), and going looking for the oopses that using VC++ will let creep into your code.)- it's not quite as hard as one would think it is.
      correct me if im wrong but did you just say it isnt difficult to plug opengl in?

      Comment


      • #33
        Cedega is shit. Don't buy that crap. Everything Cedega does, Wine does better.

        But nothing beats native ports, when they are done around the same time as the Windows port.

        UT3 is not a great game. It won't last. It's not played that much anymore. It's sad. I used to look forward to it.

        Comment


        • #34
          Originally posted by L33F3R View Post
          correct me if im wrong but did you just say it isnt difficult to plug opengl in?
          It's not simple- it's just not as difficult as people repeatedly keep making it out to be.

          There's some translation of behaviors to do (because there's not always a 1:1 mapping of what DirectX does to what OpenGL does- it's close, but not 1:1 in about 1/3rd the cases...) but it's not this herculean effort in most cases to move one in over the other and many game engines abstract out the rendering/input/sound/etc. stages so you end up pulling the plug on portions. The only really, really annoying as hell, can't get it good enough to satisfy people, part of DirectX isn't the 3D part- it's the network play part. And it's not so much the code itself- they're doing reasonably easy to replace things (I know, I've done it twice so far...), but you can't reproduce the wireline if they use DirectPlay. Hence the incompatibility in multiplayer with Ballistics, the hopefully soon to be arriving Bandits: Phoenix Rising, and many of the other recent ports. Only if they roll their own or use something like RakNet or (ugh...ugh...ugh...) GameSpy, can they have wireline compatibility with Linux and Windows.

          Comment


          • #35
            Originally posted by Remco View Post
            UT3 is not a great game. It won't last. It's not played that much anymore. It's sad. I used to look forward to it.

            It's sad, really. This all is being made to reflect poorly on Ryan's work. I'm pretty sure he's got it "done" much like Prey was done for a LONG time (I can't say HOW long- covered by NDA...) and was held back for non technical reasons. The game was largely ready- but something, probably like with UT3, was holding it back. In the case of Prey, it's "okay" it came out later. In the case of UT3...I'd stick a fork in it, it's done.

            Comment


            • #36
              Originally posted by Svartalf View Post
              [...]but you can't reproduce the wireline if they use DirectPlay. Hence the incompatibility in multiplayer with Ballistics, the hopefully soon to be arriving Bandits: Phoenix Rising, and many of the other recent ports. Only if they roll their own or use something like RakNet or (ugh...ugh...ugh...) GameSpy, can they have wireline compatibility with Linux and Windows.
              Maybe it would be worthwhile then to use (and improve) Wine's implementation of DirectPlay.

              Comment


              • #37
                Louise:

                I agree with you in that Hurrican is probably the best freeware game ever made (for Windows) and is without doubt the best ever freeware (Turrican) remake I've ever seen. However, Hurrican has worked 100% under normal wine for some time so theres no need for Cedega if thats all you really want to play. Additionally, the Hurrican developer is very keen to see it get ported to Linux but he doesn't have the time or know-how to do it himself. I brought this up on the openpandora forums and a few people expressed interest in porting it but I don't know if such work is really under way now or not.

                You know what I'm absolutely f$%kin sick of all these totally dumb "Year of Desktop Linux- Never gonna happen - there are no games" articles- multiple new ones appear on the tech news sites every bloody day. I would love to see the editors of tech news sites ban "Year of Desktop Linux" because they should realise that cannot, will not happen. In another 10 or 15 years we'll be able to look back and go "Ah yes- 2007 through to 2017 (or whatever period) was the 'Decade of desktop Linux'". We all know the many reasons why Linux is technically and ethically superior to proprietary OSs but it isn't these issues that stop a large scale conversion happening, instead its legacy windows apps, hardware that was never supported under any OS other than Windows, people liking what they know, laziness, Redmonds' dodgy business practices etc etc.

                As for Linux games- I'm very encouraged to see stuff like Unigine and gallium3d popping over the horizon and it certainly helps that AMD/ATi take Linux a lot more seriously than they did a few years ago, what with AMD including Linux drivers in the box on day of release for their latest and greatest cards now, open hardware docs etc.

                People often cite game development as an important area of software that isn't going to flourish using the Linux dev model- big games require equally big funding to pay all the artists,designers and programmers required to work full time to produce. A lot of the best FOSS games are based upon tweaked commercial engines that became open sourced, hence very few commercial quality games (if any) have been produced from scratch using the FLOSS dev model.

                As Linux ever increases in popularity, there is of course a steadily increasing demand for Linux native games. We can see from the Phoronix forums that there are indeed a considerable number of Linux users interested in and willing to pay for good, Linux-native games. Most Linux users do NOT expect every single piece of software to be free and most do not share Stallmans views on closed-source, non-free software being totally evil and to be avoided like the plague.

                However, the real problem is that commercial/closed source games don't sit well upon the ever evolving, ever shifting and changing mass of libraries, programs and APIs that make up the different distros. As an example, my favourite commercial Linux games are Spacetripper and Mutant Storm by pompom games. Although these games were only released a few years ago, they refuse to run on any Linux distro from the past couple of years due to a glibc incompatibility. If I want to play these games under a recent distro then my only option is to play the Windows versions under wine, which is a sad state of affairs.

                Seeing as 99.9% of the big, popular Linux apps are 100% FOSS, this constantly moving dev target isn't really a prob because everything just gets recompiled, repackaged and/or patched where necessary. So, unless circumstances drastically change and the whole Linux ecosystem 'stabilises' commercial game developers would seem to have two options:

                1) Release packages for all the major distros, updating them for new distro releases AT LEAST for the shelf life of the game. Obviously, this is a lot of extra work for the game devs and we still have the problem of 'bit-rot' stopping these packages working under later revisions of specific OSs, unless everything is statically linked in which case I suppose the game might as well be on its own live CD, chroot install or whatever.

                2) Much nicer as it pretty much solves all the probs in 1 but much less likely to happen would be if game studios wanting to support Linux would 'open source' (most likely under a restrictive, non FSF approved license) the game code so that if the provided packages failed to run, the user would at least be able to get the game running by compiling it from scratch. All the data files would of course remain proprietary and copyrighted.

                So yes, the lack of games under Linux is certainly not just because of a relatively small market- there are many important obsticles that need to be overcome before Linux gaming can really take off.

                Comment


                • #38
                  in response to that. Fallowing the ubuntu strategy of LTS they could keep major compatibility while making it more closed source friendly as developers could target a specific long term kernel version. Let the distros deal with picking what kernel to use as usual.


                  just my 2 cents.

                  Comment


                  • #39
                    Originally posted by danboid View Post
                    Louise:

                    I agree with you in that Hurrican is probably the best freeware game ever made (for Windows) and is without doubt the best ever freeware (Turrican) remake I've ever seen. However, Hurrican has worked 100% under normal wine for some time so theres no need for Cedega if thats all you really want to play. Additionally, the Hurrican developer is very keen to see it get ported to Linux but he doesn't have the time or know-how to do it himself. I brought this up on the openpandora forums and a few people expressed interest in porting it but I don't know if such work is really under way now or not.

                    You know what I'm absolutely f$%kin sick of all these totally dumb "Year of Desktop Linux- Never gonna happen - there are no games" articles- multiple new ones appear on the tech news sites every bloody day. I would love to see the editors of tech news sites ban "Year of Desktop Linux" because they should realise that cannot, will not happen. In another 10 or 15 years we'll be able to look back and go "Ah yes- 2007 through to 2017 (or whatever period) was the 'Decade of desktop Linux'". We all know the many reasons why Linux is technically and ethically superior to proprietary OSs but it isn't these issues that stop a large scale conversion happening, instead its legacy windows apps, hardware that was never supported under any OS other than Windows, people liking what they know, laziness, Redmonds' dodgy business practices etc etc.

                    As for Linux games- I'm very encouraged to see stuff like Unigine and gallium3d popping over the horizon and it certainly helps that AMD/ATi take Linux a lot more seriously than they did a few years ago, what with AMD including Linux drivers in the box on day of release for their latest and greatest cards now, open hardware docs etc.

                    People often cite game development as an important area of software that isn't going to flourish using the Linux dev model- big games require equally big funding to pay all the artists,designers and programmers required to work full time to produce. A lot of the best FOSS games are based upon tweaked commercial engines that became open sourced, hence very few commercial quality games (if any) have been produced from scratch using the FLOSS dev model.

                    As Linux ever increases in popularity, there is of course a steadily increasing demand for Linux native games. We can see from the Phoronix forums that there are indeed a considerable number of Linux users interested in and willing to pay for good, Linux-native games. Most Linux users do NOT expect every single piece of software to be free and most do not share Stallmans views on closed-source, non-free software being totally evil and to be avoided like the plague.

                    However, the real problem is that commercial/closed source games don't sit well upon the ever evolving, ever shifting and changing mass of libraries, programs and APIs that make up the different distros. As an example, my favourite commercial Linux games are Spacetripper and Mutant Storm by pompom games. Although these games were only released a few years ago, they refuse to run on any Linux distro from the past couple of years due to a glibc incompatibility. If I want to play these games under a recent distro then my only option is to play the Windows versions under wine, which is a sad state of affairs.

                    Seeing as 99.9% of the big, popular Linux apps are 100% FOSS, this constantly moving dev target isn't really a prob because everything just gets recompiled, repackaged and/or patched where necessary. So, unless circumstances drastically change and the whole Linux ecosystem 'stabilises' commercial game developers would seem to have two options:

                    1) Release packages for all the major distros, updating them for new distro releases AT LEAST for the shelf life of the game. Obviously, this is a lot of extra work for the game devs and we still have the problem of 'bit-rot' stopping these packages working under later revisions of specific OSs, unless everything is statically linked in which case I suppose the game might as well be on its own live CD, chroot install or whatever.

                    2) Much nicer as it pretty much solves all the probs in 1 but much less likely to happen would be if game studios wanting to support Linux would 'open source' (most likely under a restrictive, non FSF approved license) the game code so that if the provided packages failed to run, the user would at least be able to get the game running by compiling it from scratch. All the data files would of course remain proprietary and copyrighted.

                    So yes, the lack of games under Linux is certainly not just because of a relatively small market- there are many important obsticles that need to be overcome before Linux gaming can really take off.
                    Lets not also forget the other hurdles that linux faces.

                    1) What schools teach. Lets face it almost every class out there concentrates on DX. Which brings me to my next hurdle.

                    2) Lack of a single unified API for game development. DX shines there which leads to the next point

                    3) Good complete documentation that is easy to find. If you download the DX SDK it comes in a nice complete package along with excellent documentation and working examples of the newest features available. Also resources like books are rather skimpy vs the overwhelming flood of DX books available.

                    Comment


                    • #40
                      @danboid:
                      Solution is simple: make engines which "use" the Linux development model. All these problems stem from an entire aged view on engine design: one black-box type software hard to change. As long as engines stick to this black-box design it will always yield troubles with Linux. You have a dynamic OS? So make a dynamic engine! That's the only solution in the long run. The rest is temporary fixing symptoms.

                      Comment

                      Working...
                      X