All right, I''m anonymous so no one will believe me, but I counted it up and I have had a part in about 20% of the commercial games running on Linux so I think I can talk about the subject with some authority.

People saying "The problem is not technology": You''re half right. The only OpenGL drivers on Linux that don''t suck are NVidia''s. These need to get better, games or no games. UT2K4 runs on Linux so we''ll just say that *most* of the problems aren''t technical.

People saying "cost/benefit" are on the right track. The financial incentive to port to Linux is very small. Linux games sell only slightly better than gift-wrapped excrement. You may even lose money porting to Linux, as it is widely speculated that Bioware has. However Mac and Linux ports have worked out well for a company set up to do them, which Epic is. They have one employee doing all their alternative platform work (Mac and Linux).

Bootable CDs... this is no new idea. Back when I worked at Loki they had a bootable CD whipped up to play Sid Meier''s Alpha Centauri. It''s an interesting idea from a technical standpoint, but not from a practical one if you want my opinion. A lot of that has already been touched on.

A games oriented distribution... I don''t think it will change anything. Linux needs people buying games. That means the people that run Linux now need to step up and pay for games on Linux. Unfortunately the only company selling Linux games right now is Linux Game Publishing and I think it''s safe to say that they''re not getting anywhere near the hype or sales of Loki.

That quote from the Frozen Bubble page... that was interesting. It''s an example of how misguided a lot of the Linux users are. Frozen Bubble, while completely unoriginal, is a solid game, but I''d rather have people developing games for Linux that have a more sensible view of the world.

And finally, the sentence "Linux isn''t ready for the desktop" is pretty meaningless unless qualified. What in the world does "ready for the desktop" mean? Do you even know, or does it just make you sound smart? What do you think Linux needs before it''s "ready for the desktop"? Senior citizens as users?!

Good points, AP.

well... i guess the problem here is ''commercial''. linux stands front most for free distro = free apps. in fact a free os does not oppose commercial games but to be honest... if i can chose between a free game for my linux box or a buy game, i would take the free one, just because i (my personal opinion, no flaming) have got the impression over the past years that free stuff had more of this feeling a game just needs to make if great, which in my eyes got somehow lost in the commercial scene (ut good, ut2k3 OMFG; deus ex good, dxiw OMFG; hl greate, hl2 OMFG) and further more because i use self-compiled software on my system as far as possible.

i can understand very well why companies don''t venture into linux cause they fear people not buying it and the laking support.

support is a non issue. it''s solved with a few simple words: We don''t support the linux version.

the entire problem is that porting to linux doesn''t generate sales.

So buy any linux game you find that looks cool. Go buy NWN (the game I was trying to think of) and make sure they know you''re buying BECAUSE of linux support.

this would imply i like the game first... :lol:

quote: Original post by C-Junkie
support is a non issue. it''s solved with a few simple words: We don''t support the linux version.

If I was to run linux, I''d be dual booting with windows to run Win32 games anyway. So if a company comes along and says the linux version is unsupported, take a guess which version I''ll be buying.

quote: Original post by Oluseyi

quote: Original post by pkelly83
What the problem with rebooting a computer to play a game?

This isn't an NES. This is a PC. We don't reboot PCs to play each game. We didn't 10 years ago, we shouldn't now.

Run a windows game, Then another, then another. You have to reboot becuase system resouces get eaten, and I mean a real top of the range game on an average home user's PC, not solitaire.

quote: Original post by Oluseyi

quote: Original post by pkelly83
A properly configured Linux system will boot in 5-10 seconds...

Lies. The POST alone takes longer than that.

Okay maybe 30-40 seconds (To load X to the login screen), windows XP took at least 3 minutes on my PC. You can configure some Linux systems to boot faster, depending on their application domain.

quote: Original post by Oluseyi

quote: Original post by pkelly83
You're effectively turning a PC into a console.

What a waste. It would be smarter to build a dedicated system out of PC components. Wait... That's been done. Even worse, it was tried with Linux before - and failed. Twice.

I was an active member of the Indrema Developer Network (IDN) before joining GameDev, so I speak from experience. The big mistake they made at Indrema was hype over substance: cheeky press releases, 3D mockups, "developer tools" - for a machine that probably wasn't even a prototype yet. It was a disgraceful exercise in ineptitude.

You're just being overly sarcastic here. A legacy free PC would be cheaper and more versatile, but I accept your point. Even the XBOX has a version of Linux for it, and that boots straight from CD.

quote: Original post by Oluseyi

quote: Original post by pkelly83
Internet kisosks and multimedia kiosks already do this.

Orthogonal application domain. The demands of such systems are different - for instance, they don't need to reboot for each application/user.

Okay but many of these are x86 PC based. I saw one being installed on Friday. It had SuSE 8.2 on it running on some kind of Compaq PC. Being installed and setup by your average handyman not a computer engineer. Yes, you probably could train a monkey to do it too.

quote: Original post by Oluseyi

quote: Original post by pkelly83
...you could build a very powerful arcade machine this way.

Yes, you could. But arcade machines only run one game for protracted lengths of time. That is not what you are suggesting - again, orthogonal application domain.

Was that not the point of having a Linux system that boots one game? You can't say that what I'm suggesting it's an orthogonal application domain, thats what I mean't in the first place. It has been done too.

quote: Original post by Oluseyi

quote: Original post by pkelly83
I recently gave up on Windows all together, because I got sick of XP not being able to do anything useful.

Define useful. Substantiate your claim. Quite frankly, I find people who make statements like this to generally overestimate their own productivity.

I needed a webserver to test a college project IIS won't work on Windows XP Home and I don't have €497 for Pro. Windows XP won't run some of my old Windows games very well, Linux will with the WINE API Layer - and perfectly too (I know WINE isn't perfect yet). By useful I mean Linux is MORE useful to me, personally, right now. If you like win then thats your choice and I won't dispute it, but don't critise in such terms as you used above for using a different OS. If you don't like Linux why are you posting in the Linux forum?

quote: Original post by Oluseyi

quote: Original post by pkelly83
A distro like SuSE or Mandrake is desktop ready and a breeze to configure.

Obviously it must be, since you said so. Right? Right?

Yeah fine you obviously agree. Have you used them yourself? Desktop readiness, is of course, a matter of opinion.

I do accept you have more a lot more industry experience than me and I have very little, I am merely giving a few opinions and ideas about how one might go about implementing such a system. Having industry experience still does not give you the right to be condescending to people, perhaps you could give advice and substaniate your claims, like you asked me to.


[edited by - pkelly83 on March 24, 2004 12:40:05 PM]

[edited by - pkelly83 on March 24, 2004 12:41:08 PM]

As others have already pointed out, there is not much apparent value in a ''live'' cd. That being said, however, you may want to look into Gentoo, particularly the Unreal Tournament 2003 live cd that they were using to promote thier gaming-sources kernel a while back. HTH

I am surprised that no one has mentioned Quake 3 Arena in this thread. Runs great on linux, and has tons of mods. Also just out is Unreal Tournament 2004 (with Linux installer on the CD)

But here''s the thing: how high, exactly, is the cost of developing for Linux. If I were to believe the portability hype of various development tools/libraries, it should simply be a recompile and some testing - miniscule. Which means that virtually all benefit would outweigh cost (so long as no investment is made in shelf space or delivery). But it appears not to. Can you explain that to me?

Unprovoked barbs from Linux-hostile moderators notwithstanding, why use *any* valuable development/support/testing time porting to other platforms when the original platform targets 97% of the gaming market? (I just made up that number). It is a cycle; no new games -> no new gamers -> no new games.

Regarding cross-platform libraries: while recompiling, testing, and supporting additional binaries may cost a ''miniscule'' amount (and that is debatable), retraining developers and working at below maximum efficiency during this process does not. SDL is a stable, efficient, intuitive library, but how many DirectX programmers would be happy to find out they have to start learning/using SDL instead?

It''s simple, really. Linux, while growing, is still a relatively small market. Partially this is because Win32 comes preinstalled on nearly every PC produced (another issue altogether). The actual market is also somewhat undervalued because of developers (id, for example) who sell Win32 only cds at most retail outlets and have a downloadable linux installer. My Win32 Q3A and Team Arena cds, for example, never got installed on Windows. While I don''t think my situation is particularly prevalent, more and more game developers will be forced to target Linux as the market grows. This is a natural shift that will take time.

quote: Original post by grazer
But here''s the thing: how high, exactly, is the cost of developing for Linux. If I were to believe the portability hype of various development tools/libraries, it should simply be a recompile and some testing - miniscule.

First, don''t believe the hype. Porting is usually never as simple as a recompile. A generally portable codebase may still take two people a week to port. That''s two programmer salaries, for a week of work. It''s highly likely that you won''t be able to grab two programmers that have never used Linux before and start them on a port and expect it to take a week. Think months. If you have to license extra libraries, that''s more cash involved. Porting costs are minimized by enforcing good practice throughout development. As I mentioned above, Epic has one programmer doing all of their alternative platforms. The overhead is generally low for them because they keep the engine as portable as possible through the whole development process.

quote: Regarding cross-platform libraries: while recompiling, testing, and supporting additional binaries may cost a ''miniscule'' amount (and that is debatable), retraining developers and working at below maximum efficiency during this process does not. SDL is a stable, efficient, intuitive library, but how many DirectX programmers would be happy to find out they have to start learning/using SDL instead?

Many people will be hesitant to use SDL because of its license. The issue is not even SDL vs DirectDraw anymore, it''s OpenGL vs Direct3D. Direct3D, at least currently, is the clear winner for high-end, commercial PC games.

AP, your opinion is obviously held by the majority of the games industry. But it's just not correct. It took me half a week of around 4 hours a night to get my game engine running in windows and linux, just recompile and it works.

Truth is, I'm not that good. Certainly not as good as pros who've been in the industry for years. So how do you explain it? It was pretty easy for me.

I agree porting is much harder if you don't keep cross platform compatibility in mind throughout. But that doesn't involve any extra work, just a change in mindset.

It isn't about crossplatform stuff being harder to use, or taking longer. It's about lazy people not wanting to learn new ways to do things. Maybe I've had a bonus here in that I've generally tried to learn a crossplatform way to do anything I've learned, and so haven't got into a mindset where I see it as 'hard'.

Your claim abour d3d: if you mean winner in terms of popularity, then yes. Anything else, then no. OpenGL is technologically on a par with d3d, but is more flexible. This is where you bring up the 'extension hell' yes? Nothing that can't be sorted with one of many libs findable thru google, and 10 mins to get it working.

edit: And the great thing is, now I've done it I've got a base that all my games can build on, that works on multiple platforms. I'll never have to write it again. Now I've done it I have an extra bunch of people to sell my game to (were I a pro) essentially for free.


[edited by - mrbastard on March 24, 2004 3:08:54 PM]