Is the Key to Linux a Games-Based Distro?

An anonymous reader writes "If in the FOSS community we could only get our act together and launch a game-based distro, we will be home and dry. That, at least, is the view of one British games enthusiast, Ian Bonham, who says in the short Linux World article: 'I would be happy to help a group of volunteers create a distro based on games, because I believe that's where the next generation is - NOT in giving away copies of Linux or OOo. That's a short-term ideal. The PS2 and the X-Box(sic) run Linux, so let's create a distro that turns home PC into a console with development potential. Expand that distro to the consoles. And lets get some 'killer' games on that disk.'"

Comments

Woo

[Anonymous Coward]

Can we get Tux Racer? Now that's livin'...

Re:Woo

[somethinghollow]

Hopefully X-Bill doesn't count as a killer game. Linux gaming is getting better (Quake 3, and other OpenGL based games don't requier much re-working to port to other platforms, AFAIK), but let's not kid ourselves. Games that come with window managers usually just can't be touted as features. At least Microsoft never said (to my knowledge) "Our OS comes with games built in," referring to Solitare and Minesweeper (and whatever else comes with XP now).

Re:Woo

[mahdi13]

Isn't the Linux version of UT 2004 included in the box?

YES

Re:Woo

[shepd]

>At least Microsoft never said (to my knowledge) "Our OS comes with games built in," referring to Solitare and Minesweeper (and whatever else comes with XP now).

Wanna bet? ;-)

Re:Woo

[Pxtl]

Heheh, this of course highlights the problem with Linux gaming that we all refuse to admit:

they suck.

I'm sorry, but a million clones of warcraft II and Quake 1 does not a gaming environment make. Still, there are Tux games that have real futures. I'm a windoze user and I can think of a few free software games that I play incessantly. Now, as I understand it, Tux Racer is not multiplayer. At this point, I stop giving a shit. There are some good ones like Cube and Armagetron, but even they are only skeletons of games - they have the minimum "get online and play" gameplay, and graphics that would be current for 1997. Still, I love them to death and have sunk countles hours into Cube (wouter.fov120.com).

The fact is that there is not a complete free software offering to counter the Quake and Unreal engines. Yes, crystalspace is nice, but it just doesn't have the complete feature-set and complete game to build a model of a full game onto.

Think about this - all of the retail engines have heaps upon heaps of mods that a) completely replace all of the in-game media and b) replace tons of code. Linux does not have a similar free alternative to these frameworks. As such, people that would like to develop for a free platform are instead relegated to retail world, and games that could become the basis for a free software community stay fringe.

Look at the best offerings of the free software community for gaming engines - CrystalSpace and various flavours of the Quake 1 and 2 engines - tell me that they really come close to the Unreal or Quake 3 engines, much less the current generation.

now, one thing I can't help but notice - free software games do not seem to be aware that I own a joystick, much less many joysticks. People who talk about "linux as a console" seem to neglect this little detail. I have a windows 98 box and an old gravis multiport wired to my TV set, and I have a handful of games that I play on that. The PC selection for games that support multiple joysticks for multiple players on a single screen is damn small, and all of them are DirectX-based games (blaster disaster rulz). None of the SDL-based offerings have shown me anything in that department.

Take the Quake II engine, give it a non-shitty modeling system, some physics, and some real shader support and convert it over to a Python or some other script-based framework so people can develop for it easily. Then re-implement a basic online CTF+DM game for people to start their work from. Then, maybe, Linux games will be able to compete. I haven't seen anyone succeeding at that. Even Doom engine ports are still painfully primitive in terms of script support and other features you'd expect them to get after so long.

Re:Woo

[abe ferlman]

Games that I have played in the last couple months on my linux box:

Neverwinter Nights and expansions
Quake 1,2,3 and mods
Return to Castle Wolfenstein
Unreal, Unreal Tournament, Unreal 2k3/2k4demo and mods
Warcraft III and expansion
Diablo I and II and expansion
Jedi Knight II: Jedi Outcast
Angband :)
Descent III
Half life and mods

Games that I've been really annoyed that I couldn't play on my linux box in the last few months:

Unreal II
Various midi-enabled piano-tutor games

Linux gaming is not where it was a few years ago. You might need a winex rpm to get past the copy controls for some of the games but it's cheaper and better than buying windows (or any proprietary console or whatever).

Re:Woo

[savuporo]

The trouble with all this is, nowadays, an engine ( or a programmer ) does not a game make.
In a recent thread over at Beyond3D forums, we asked developer of Max Payne 2, how are the costs distributed in a project like this between middleware, code development and content development. He said:

• middleware 10%
• content 60%
• and the rest is code, i.e. only 30%

So, unless you wire up some really innovative procedural content generation routines, Open Source Software alone isnt going to get you far in game development.
Btw, im quite certain that the trend is ever growing, i.e. content part is going to take up more and more of game budgets.

Re:Woo

[Pxtl]

My point is that independant modder communities have been making content for free for over a decade. How many total conversions are there for Half Life and Quake 3, all for free? Those peopel replace teh base game content so completely that they really are just useing the codebase. What the OS community needs to do is make a platform attractive enough to bring them in. It would be attractive for them too - their game would be a free standalone instead of a mod for an existing game.

www.moddb.com will blow your mind with the amount of projects under way - most die early, but alarmingly many run to completion making a full game from nothing but an FPS codebase, replacing all other content and adding all relevant code. If ever there was a community that _needed_ to harvest this power, it was OS.

We need an OS answer to Half-Life. Not interms of plot or gameplay, but interms of the mod community around it. Then you'd have enough games on Linux that no-one would ever complain about Linux not being a game platform.

Re:Woo

[Eshock]

Tenebrae is a great example of a free open-source game engine with linux support. It even supports pixel shaders and 3d audio.

Interesting

[Ummagumma]

I agree with this assessmanent, however, one of the biggest challenges is to get peoples legacy Windows games to work, which is quite the challenge, if possible at all, on a reliable basis.

Re:Interesting

[scumbucket]

Isn't the challenge not to make windows games run under linux, aka wine, but to get game publishers to release linux versions of their games?

Now a standard linux distro aimed solely at game developers to make their life easier might be a better way to go......

Re:Interesting

[oskillator]

I agree with this assessmanent, however, one of the biggest challenges is to get peoples legacy Windows games to work, which is quite the challenge, if possible at all, on a reliable basis.

Running legacy DOS games natively is a pain when it's not impossible, but the DosBox emulator does a really good job at it, and there are builds available for all major operating systems.

Re:Interesting

[SillyNickName4me]

A bootable, playable CD would solve a lot of headaches for game developers.. provided you can solve the driver issues.

Re:Interesting

[Anonymous Crowhead]

Bah, I do that now. I reboot into Windows whenever I want to game. That is the sole purpose of my dual boot.

Re:Interesting

[mdfst13]

The problem is that if someone has a game that they really like (the Warcraft 2 Expansion Set comes to mind for me), then they will want to be able to run that game on their PC. If they can't, it doesn't really matter if the PC runs newer games (which is what manufacturers will port) or not. If they can't run their existing favorite games, they can't run Linux full time.

For this reason, I consider a games distro to be one of the worst ideas to gain Linux converts. There are just too many games.

Concentrating on general (email, web browser, word processing, spreadsheet, etc.) and specific (CAD, web design: e.g. Dreamweaver) applications makes more sense. There the issues are more in terms of supporting a few apps that someone uses almost exclusively. File compatibility is the important part, not application compatibility (I don't need to run Microsoft Word if OpenOffice can load and save .doc format; Evolution can connect to my Microsoft Exchange server; etc.).

I especially like CAD as a Linux app, because CAD designers frequently run *only* their CAD software on their PC. Even if they can't run any other software on it, it doesn't matter. They wouldn't anyway. Further, CAD uses gobs of resources and is thus better suited for lean running Linux (system processes leave more room for CAD processes).

IMO, games should be one of the last areas of focus for Linux developers. There are just too many legacy games which will never get ported. Thus promoting hacks like WINE. Linux should concentrate on its own apps, not pretending to be Microsoft Windows.

In the meantime, consider looking at multi-platform game development engines like those provided by Garage Games: http://www.garagegames.com/pg/browse.php?type=deve lopment

Multi-platform engines enable game designers to get both markets easily. Ideally, they could develop on Linux (less system process bloat means faster compiling) and test the game on Microsoft Windows.

Btw, now that I have actually RTFA, I notice that the author is talking about something like bootable CDs with games on them. This already existed: that's exactly what Gentoo Games CDs were. The website ( www.gentoogames.com ) no longer seems to work, so I'm guessing that it never took off. Morphix also works on this (game specific live CD).

Another reason not to wait for games is the problem of too much of them being content rather than code. Modern games are frequently based on impressive 3D graphics (content) and movies (content) rather than spectacular game engines (code). Several of the big time multi-player games have already released Linux version (e.g. Id Software products). Until Linux has much more of a market share (at least 20%), we can't expect anyone to develop a Linux only game.

Look at how much money Microsoft is losing on XBox. Not a problem for them, since they have the money to lose. Linux doesn't have those kinds of resources. One game wouldn't do it. To really draw people would take ten or twenty.

Re:Interesting

[It'sYerMam]

I disagree - it's not *just* legacy games that people want. Of course, porting things such as UT provided a boost in the market - I play UT on Linux.

I think there're a couple of games that need to be playable on Linux - UT, Half Life, C&C, Max Payne...
The "classics," basically. If you're a die hard fan of Billy Bob's pro Stamp Collector, then I doubt you will convert to Linux, but there're certain games that would get people thinking - when coupled with a few new games released for Linux.
It may be possible to boost confidence by getting programmers to port games to Linux as quickly as possible, releasing the ports as near to the release date of the Windows version as possible. This would then convince people that new games are coming out for linux, hopefully converting some, meaning that the professional developers would consider releasing a linux version as well, not just 3rd party programmers.

Personally, I'd be happy on Linux with UT, Hitman 2, and the promise of Thief 3 and Half Life 2.

Re:Interesting

[cgenman]

The PS2 and the X-Box(sic) run Linux, so let's create a distro that turns home PC into a console with development potential.

There's a man who has really thought this one out.

We need some "killer'"games on the CD.
We need the source for the games on that CD.
We need that CD in places like Electronics Boutique and GAME.
We need kids able to pick up that CD (or DVD, with respect to another learned friend posting here) and turn their PC into a games console, without ruining Mum's or Dad's official documents.

Ok, to sell this as a platform, you have to add the words "exclusive" after the word "killer." Otherwise, you just have a platform that can play the games that are already available on Windows, and there is no incentive to switch. But making a "killer" exclusive game requires more than just 80 hour weeks and a 10 million dollar budget: it requires both of those things many times over to create a single "successful" title. A "killer" title might require 30 or 40 fully-funded projects that reach the store. If effort was enough, we would have 100 "killer" titles every year.

Good luck with source. If you though cheating in online games was rampant before...

Besides, most videogames don't lend themselves very well to open sourcing. The industry just moves too rapidly, and games aren't something you're going to improve because you use it every day. There is, of course, NetHack and other Open Source games that do incredible things. But let's be real here, would you buy a box with NetHack on the cover if it was sitting next to a box of Doom 3?

Getting on the shelf in E.B. is not that difficult once you have actual street cred and some cash to back it up. E.B. loves cash. But as this seems to be lacking a business model (or, for that matter, a plan), I don't know where they would get either.

As for transitioning to consoles... That doesn't make any sense. If the Phantom and ApeXtreme are such bad ideas, why would a Linux based ApeXtreme be any better? Why do you need a console when you can have a computer with TV out and hit the mass market? Or, conversely, why would the average person want to run Linux on the PS2?

He fails to mention that the CD would need to be bootable, ALA Knoppix, or else the formatting process would "ruin Mum's or Dad's official documents." Because, as we all know, official documents require Rockin' graphics cards left in public spaces or they get lonely. Likewise, you will need to be able to install to disk, like Knoppix, or else there can be no platform transition. You need to support a large amount of hardware, like Knoppix, and have a lot of available games, like Knoppix. Oh, and you want it based upon the most solid binary distro available with the clearest licensing, like Knoppix. Are you seeing where I'm going with this?

No. What he really should be doing is going to game development companies and pushing the idea of entirely self-contained games running on Linux. It would be significantly harder to cheat in a MMPORPG game if it ran as its own OS, booting without a HDD, and then you could offload the action processing to the individual clients without fear of modification. Lag would be a thing of the past, and MMP twitch games could be released. Ask for a hash key of random length of the CD every now and then, and you would have a very tough nut to crack. And if people did crack it by learning to hack through Linux, all the better for the platform. He could also push Linux to Sony and Nintendo as a way to quickly create a solid development system for next-gen gaming. Unlike Windows, Linux's multiprocessor kung-fu is superior, and would probably like the Ps3's 18 processor architecture in a way that nobody else would. It might even make it a bearable system to work on.

In short, this guy doesn't have a firm grasp on the industry. It would be great to push Linux to the people who control the standards, but pushing the OS without codifying it into the gaming ecosystem somehow is suicide. At least Sisyphus got near the top of the hill before the boulder rolled back down.

Re:Interesting

[Kethinov]

I was going to moderate this discussion, but somone has to say this:

A standardized Linux distro is needed. Not another obscure niched one except built just for games. Because a standardized distro would be inherrently best for games.

Imagine if Bruce Perens' UserLinux came with nvidia and ati's binary drivers and automatically installed them during the distro installation. Currently no distro that I know of does this, the drivers must be manually installed.

One could argue that in most cases you have to do the same thing in Windows, but in Windows all that requires is double clicking an install file. In Linux you have to usually exit X, check dependencies, and all kinds of other cryptic stuff.

Finally, the one thing that we most need that a standardized distro can provide, is a standardized directory layout. None of this /usr/bin or /bin or /var/usr/bin confusion. If one distro took over by having all the features that desktop Linux needed, which in my view is basically Fedora to unify toolkit look across gtk, gtk2, and qt, but with better hardware detection (ala binary non OSS drivers) and better package management (ala automatically installed apt-get), the standardized directory layout would encourage more Linux ports of games.

As a software author, most authors only release their software as source when dealing with Linux, because it's the only way to ensure that it will work in every distro. But if there was a standardized directory layout and package management system, every dependency could always be found in the same spot and there'd be no need for third party package management and binary compilation.

This may seem like nitpicking, but many companies don't port their games to Linux on the sole basis that they 1. don't want to release source and 2. don't want to take the time to write an installer which can accomodate every distro's different package management, directory layout, and dependency tree.

So that, my friends, is what Linux needs. Create that and gaming will follow.

Re:Interesting

[Romeozulu]

Any game that has spare cycles isn't worth playing :-)

Re:Interesting

[mahdi13]

Disapointing? How so? It is identical to the Windows version...there are zero differences.
If you were disappointed with the 'Linux' version, most likely you were disappointed with the game in general...

Re:Interesting

[mattyrobinson69]

we dont need one official distro. we need a standard base for the distro's to base themselves on. that way people still get to choose a distro, but the distro's would be more similar. Mandrake could still work on simplifying stuff, slackware could still be unix like, gentoo could still be for die-hards, etc. is anybody working on this at the moment? what is Linux Standard Base - is that what im talking about?

Re:Interesting

[Kor49]

I agree wholeheartedly. This is not stricly related to games, but it is critical.

You just need to get out of the "if you want something, compile it yourself" mentality. I am a software developer myself, but I hate downloading source code that I have no interest in reading. I hate looking for Mandrake RPM's on the net, too. I hate when RPM's require other RPM's. I just wanna be able to download whatever binary and run it as soon as the download finishes. And no, I don't even wanna know about apt-get, rpm, or whatever else is the proper tool.

Just like it's in the Windows world, when I click on an application's Setup.exe, it should just install. I don't care if you'll have to statically link everything, or implement another scheme.

In the OSS world, the itch that gets scratched is the one that the developer has. This is the itch that belongs to people who either don't have the time or the talent to solve it.

Re:Interesting

[kundor]

that's really strange.

I get identical performance, including framerates, in linux and windows with both ut2003 and the ut2004 demo.

I also play Alpha Centauri, Neverwinter Nights, Enemy Territory, Kohan, and Civilization: Call to Power very satisfactorily in linux, thanks largely to tuxgames.com.

The only reason I still have a dual-boot is Rise of Nations. There are more games available for linux than people realize. Everything id publishes, and epic game's stuff, and an eclectic assortment of others.

Re:Interesting

[Ageless]

This is such an important point. Some friends and I were just discussing this as to why Linux still isn't ready for the desktop.

Half the time when you download some simple program you end up needing a dozen other libraries for it to run. Why the hell don't people staticly link this stuff? The APIs for many libraries are so unstable that the idea of "What if I wanna update libBlah later on?" doesn't work and it's not all that important that save on transit or hard drive space any more.

I write quite a few free programs, and I always staticly link them with everything they need. It might mean downloading an extra few hundred KB, or even a few MB but in the end the user is not put out of the way and it "just works". As the developer of the program I know what version of what my program needs, and I am more qualified than any one else to determine that. It should be my responsibility that my program includes it.

Re:Interesting

[dasmegabyte]

Really?

Because I seem to remember the days of the Commordore 64, the Apple ][ and the PC Jr, when EVERY game was on a bootable disk.

And I remember being very happy when DOS reared its mighty head, and I no longer had to reboot my machine to start a new program.

Why should we take a step back in time JUST so developers can have a Linux based platform when they already have a ubiquitous development environment in the Windows PC?

Games Based Distro

[stratjakt]

Yeah, whatever.

There's so much missing structurally for that to even be considered. You know, silly stuff like reliable, robust video and sound drivers.

Cart before the horse.

Re:Games Based Distro

[bonch]

Not to mention this has already been done before. Heck, Gentoo provides a "games-tailored" kernel for emerging.

It's really true, there are some fundamental issues that need to be resolved before having a games-based distro. Right now, there wouldn't be that many games to play on it anyway.

Re:Games Based Distro

[Anonymous Coward]

Ever heard of a USB HID device? They're all supported.

As for multiple monitor support, it's called Xinerama and it works, as well.

Re:Games Based Distro

[Minna Kirai]

You know, silly stuff like reliable, robust video and sound drivers.

It's funny, but Linux is in much better shape for video drivers than audio ones. Since the game-capable graphics market only includes two companies, Linux is already adequately usable.

But since soundcards are technically easier to make, there's many more brands still in active use. Many gamers who buy the latest NVidias to squeeze a few more FPS or pixels might still be satisfied using motherboard audio output, or a $2.50 PCI soundcard.

Linux audio support is close to adequate... but unfortunately, the Alsa Project's longstanding philosophical refusal to move software mixing into the central driver means you still can't expect Linux to run games on any random piece of desktop PC hardware.

Re:Games Based Distro

[LittleBigLui]

Gentoo is hardly a very easy distro.

Actually it is a metadistro, hence a potential starting point for a [games-oriented] distro.

Re:Games Based Distro

[proj_2501]

software mixing is JACK's job.

Re:Games Based Distro

[Frnknstn]

Ever heard of a legacy device? They are mostly not supported.

I do think the games distro is a good idea. More and more, people are starting to want the ease-of-use of a console, but it will not be easy to pull off, especially since there is yet to be a distro with the ease-of-use of Windows.

Re:Games Based Distro

[Fnord]

Well, if you take OpenGL, OpenAL, and SDL (which wraps and integrates the other two in a portable, window system agnostic way), you get just about everything that DirectX has, I think. Except for networking, but I know alot of developers that, although they like DirectX, consider DirectPlay a monstrosity and avoid it like the plague (and honestly, how many games outside of MS Game Studios actually use it?). Is there some other part of DirectX that isn't covered that I'm missing?

Re:Games Based Distro

[TandyMasterControl]

The structural problems include lack of backwards compatibility provisions in glibc versions and higher level libraries. This is a basic feature/flaw of Linux and makes the porting of applications to the Linux platform very problematic. Nobody seems to give a damn if software published last year will run this year. The unspoken assumption is it will be recompiled or if that isn't enough, it will be rewritten. I would say this is even more pernicious as a factor in retarding the Linux desktop than the sound driver problem. Everything library-wise is always changing. If you code for profit like game creators do, instead of for the hell of it, constant change without backwards compatibility is prohibitive. And it isn't much better from the paying customer's perspective either -- what's the point of having a "games distro" if the games which you paid money for are going to break just 6 - 9 months from the day you bought the distribution cd ?

Anybody have any Loki games that still work ? I don't !

Re:Games Based Distro

[FooBarWidget]

Actually, we at the Autopackage project recognize this problem and have already made a solution. See apbuild.

Bootable Americas Army CD

[niko9]

How about a variation of a bootable Linux Game CD that you can also install later ala Knoppix?

Re:Bootable Americas Army CD

[Spamlent Green]

There's a Morphix module (Morphix Gamer) that already provides this. Plus there are related projects like KnoppixMame.

However, I tried Morphix-Gamer a week or two back and feel compelled to point out that easily half of the games that came included either would not run at all (at least not from the CD), or were unbearably slow and clunky (TuxRacer for one). My machine at home is nothing to brag about (Athlon XP 1800, 1/2 gb ram), but it seemed to me it should have been sufficient to run whatever was bundled with the gamer module.

Not sure why they bothered included games that wouldn't run tolerably well from the CD. Sure someone might choose to install it to HD, but the whole point of a live-cd is just that.

Didn't work for OS/2

[PieEye]

I was a member of TeamOS/2 and we all thought that StarDock was going to help get the OS recognized. Hah.

Of course, you couldn't just run OS/2 off of a CD with no install, and video was next to impossible to configure correctly when you didn't specifically know what video card was in the box, and networking didn't work, yada, yada, yada...

Anyway, it would certainly help to have a WIDE VARIETY of games, that rivalled ones on other platforms, etc.

Re:Didn't work for OS/2

[Brandybuck]

I was a TeamOS/2 guy myself. OS/2 faced two major hurdles in gaining mass acceptance, and unfortunately, they're the same hurdles that Linux is facing today.

1) The vendors don't have a clue. "We want to be a desktop distro. No wait! A server distro. No wait! An "enterprise" distro. No wait! We need a one-click installer. No wait! We need a remote installer. No wait...

2) Windows emulation. No one bothered to write OS/2 applications because native Windows applications ran just fine under it. Then Microsoft changed the APIs, and OS/2 finally sunk under the frigid waters. Why should I run my applications under Linux/WineX when I can run them under Windows?

3) Arrogant advocacy. This is the worst one. OS/2 died in part because most people in Team OS/2 were assholes. Linux advocates are no less impolite. Face it, no matter how much you argue the point, the average consumer will NEVER believe that Linux is the holy salvation of mankind. Yet you still continue to argue that. "Linux? Oh yeah, that's the OS with all the arrogant jerks..."

Re:Didn't work for OS/2

[Anonymous Coward]

...OS/2 faced two major hurdles...
1)...
2)...
3)...

Heh, no wonder OS/2 never gained mass acceptance.

Re:Didn't work for OS/2

[kin_korn_karn]

dead on about #3. I loved OS/2 and hated MS. I was an asshole about it and I regret that. It makes me cringe when I see Linux guys talk shit about Windows, because they are missing the entire point.

Not killer games, but killer game compatibility

[jjhlk]

It takes companies years, millions, and hundreds of megabytes to create successful games, and the success to linux is a game that actually runs on linux? No, I say linux needs to be able to run PC games (well and without hassle).

Re:Not killer games, but killer game compatibility

[October_30th]

Indeed.

I'd like to see the day when the free software community can provide not only the code for a smashing game engine (come on, surely you can replicate the unreal!), but also the work of convincing voice actors, motion-capture animators, map designers, plot authors that could write a reasonably well selling book and patient texture designers.

Killer games require megabucks.

I'm still waiting for my free Ghost Recon game.

Key

[Reducer2001]

I don't think the key to Linux will be a games based distro. The key will be my mom being able to plug in her digital camera and having all the picutres show up in a window. We can still have the command-line, but the GUI has to 'just work' with everything else on the system like Mac OS and Windows XP's mostly do.

Re:Key

[Bingo Foo]

I thought that figuring out how to make your peripherals work was what made most Linuxes "games-based" distros in the first place.

Re:Key

[Pedrito]

I don't think the key to Linux will be a games based distro. The key will be my mom being able to plug in her digital camera and having all the picutres show up in a window.

I couldn't agree more. Linux needs to be MUCH more user friendly. It needs to be much more intuitive. And I don't just mean the OS. I mean all the apps you get as well. Fortunately, there's been a good deal of progress in this area and over the last few years, Linux has improved dramatically, but it's still way behind Windows and Macs in terms of ease of use for your average technophobe.

Games? Why? If games are what draws people to a system, then people are going to buy game consoles. That's why game consoles sell so well. If people want a computer, then make them want Linux by doing the above.

The author's idea is that we should get people using Linux so that people are using Linux. So let's come up with whatever cheesy plan it takesto get them to use it. At least that's the idea I get from it. I think that's stupid. Make people want to use Linux by making Linux the best alternative to Windows for more people. Then you're on to something.

Re:Key

[Sleepy]

> I don't think the key to Linux will be a games based distro. The key will be my mom being able to plug in her digital camera and having all the picutres show up in a window.

Lets not compare Windows XP to RedHat 6.2 shall we?

Digital cameras work fine. Find a valid example. Most people dismiss Linux because:
a) Windows came with their computer. They already paid for it. WHat's the point??
b) Lack of warez for Linux. A shamefully low percentage of Windows users have totally-legal software installs.
c) usability DOES factor in, but the average person just needs a Lindows-like PC.. email,. web, office app, and oh yeah support for USB cameras and pen drives. Linux does that with great ease of use.

I can't see involving the "command line" in any of those activities... not anymore than the same job requiring regedit.exe use on Windows.

Not that I'm saying Linux is as easy for mom as XP (it's not... but it's not a huge leap).

Re:Key

[Azghoul]

Okay, I don't disagree with your apparent point that Linux needs to be easier to user for technophobes...

But I think you are cracked off your rocker if you think Windows is EASY for technophobes. Have you even SEEN 'phobes trying to use a Windows machine?

Just over the weekend I was flipping through a CD full of digital photos for a couple family members. Any OS can handle this easily, but I happened to have to use WinXP (because they're multimedia POS machine had XP on it). Double-click the first image, and it almost automatically starts up a slide show.

Pretty simple right? The amazement of said family members was depressing. They had NO IDEA what I was doing; it might as well have been magic.

In the end, I figure you can only take "ease-of-use" so far. At some point you have to say "yep, it's easy enough" and move on, because some people, NO MATTER HOW EASY YOU MAKE IT, will never figure it out.

Gnome 2.4, KDE 3.2 are both easy enough for anyone with half a brain and a few weeks of computer use under their belt. However, there are VAST NUMBERS of Americans who haven't a CLUE, and never will. They just don't care.

You will never make a computer easy enough for them to use. Never.

It will take more than just any game,

[pair-a-noyd]

It's going to take a bad ass mofo of a game, and one that's NOT available on any other platform.

Make it so attractive, so kick-ass, so awesome and so LINUX that they will flock to it.

Don't let it out for M$ and don't copy a M$ or console game.

Re:It will take more than just any game,

[Greedo]

Maybe this is what's stalling Duke Nukem Forever.

Re:It will take more than just any game,

[SoTuA]

Great! Now all we need is a company that will invest its resources in a KILLER GAME that only runs in 5% of all the desktops, and it locks out ON PURPOSE the other 90%. I can just see it:

1.- Develop a game locking out 90% of the market.

2.- ???

3.- Profit!

Re:It will take more than just any game,

[Cruciform]

Tux Racer II: Grand Theft Penguin

Nice (but unlikely)

[grub]

The Mac suffers from a shortage of games, albeit not as great as Linux, and those games sell for $$$. It's a nice thought but the reality is that you need the developers too. A whiz-bang platform without games leaves you... well... with a neat looking Linux box with a game controller.

Re:Bootable CD is the answer.

[mekkab]

I think knoppix does a great job: you can fire it up and see what it looks like, and if you want, mount a hard-drive partition for the cd, or just install onto your harddrive.

Add games and you've got teen-geek heaven.

And we will call it....

[mschoolbus]

Indrema!

I'm sorry...

[AKAImBatman]

...did I miss the point here somewhere? Just about everything I know about gaming says that the more the OS stays out of the way, the better. Now they want to replace our thin OS-like layers with a complete business/research oriented OS. Why?

Seriously, the OS doesn't *do* anything for a game. All a game really needs is a collection of APIs to transparently access low-level hardware. Threading is nice, but "green" thread libraries can be used in its stead. That's much the reason why MSDOS (save for the 640K barrier) was such a great gaming platform. The OS literally did nothing. It got the frick out of the way, and stayed there.

Re:I'm sorry...

[Minna Kirai]

Just about everything I know about gaming

Apparently you haven't heard about DirectX or OpenGL, eh?

Now they want to replace our thin OS-like layers with a complete business/research oriented OS.

Whatchew talkin bout? Microsoft(tm) Windows(r) is a "business oriented" OS; Linux has no orientation at all.

Seriously, the OS doesn't *do* anything for a game.

Exactly! Which is why Linux might (in a few years, if all goes well) be a better platform for PC gaming than Microsoft(tm) Windows!

If Microsoft continues to screw up with DirectX "upgrades" that fix one game and breaks another, then game publishers might just start shipping their installation media as bootable Linux DVDs, so their support costs can be cut away. ("Put in the disc and hold down the power button of your computer")

That's much the reason why MSDOS (save for the 640K barrier) was such a great gaming platform.

Some users might've liked it, but the programmers who had to manually support each possible piece of hardware had different opinions. Back when there were only 4 video cards and 3 soundcards, it was painful but possible. Today that the complexity of the hardware has multiplied, it's no longer an option.

Um how about not?

[tomstdenis]

That's a step backwards I think. At least in windows you can both develop/work and play games.

I think a step forward will be to get some form of standard for graphics/sound/input ala DirectX style. sure opengl, oss, sdl are all good libs but they follow the unix philosophy. That is, do one thing and do it well.

There should be a unified development tool/library that includes them all. E.g. I can install "blah" and boom I got 3d graphics, sound support, joystick/keyboard support, timers/interrupt/callback etc...

Of course that doesn't stop people from just picking their fav collection of tools [e.g. ut2k4 which runs perfectly on my Gentoo box].

Tom

What planet is this guy from?

[stratjakt]

As those games are played, kids will be encouraged to learn how they work and maybe work on their own. AMOS and Blitz basic on the Amiga formed a huge range of great games, but getting people learning C++ from an early age would lead to great things for the future, I'm sure.

Does he have any sort of clue what goes into the development of a modern "killer game"?

Programming is nothing. There are thousands of man-hours going into art assets, level design, animation, voiceover production, playtesting, etc..

The days of the kid making a neato race car game on his vic 20 are long, long gone.

And like every other twit in linux land, he offers to "help make a linux games distro, even though im not a programmer and have no appreciable skills". Which follows the standard OSS game production model:

1) Think up cool name for game
2) Open sourceforge project
3) wait for programmers and artists to come write it for you
4) ??

Re:What planet is this guy from?

[fbg111]

1) Think up cool name for game
2) Open sourceforge project
3) wait for programmers and artists to come write it for you
4) ??

You forgot the most important step:

5) Profit!

Oh, wait... (ponders F/OSS game economics)

No way

[FortKnox]

Mac has tried breaking into the PC gamining scene for decades. They even had that "bigass game thats only available on that platform" called "Marathon."

It requires two things:
Quantity of games
Quality of games

You don't need to make a gaming distro, you need a gaming distro with HUNDREDS (if not more) games already available to it. And not just net-hack and tux-racer, but big name gaming companies spitting out Linux based games.

What do you need to do this? A big-ass company with a ton of cash.

It is a proven plan. Just ask Sony how it broke apart Sega and Nintendo to get into the gaming console. Money, quantity and quality of games.

Why game systems are successful

[alfal]

The nice thing about game consoles is that all the hardware is basically the same. If I buy a game for PS2 or XBOX, I know it will work on my PS2 or XBOX. Start letting someone put the linux based game distro on any PC, and they will complain about performance and certain things not working properly because they decided to test it on that old 486 they had in the closet.

Gnibbles, KTris with over 5,000 backgrounds!

[darkCanuck]

woohoo!

My first Linux installation had me drooling at the list of games that were in the Games folder. Then, as I started each, one by one, I found the feeling similar to when you got your Burger King meal's get-the-bb-into-the-holes game.

Or, similarly, found the amazing Atari emulator only to find that those games that used to kick ass now keep your attention for about 30 seconds each - but there's 2,000 of them!

Chicken and Egg Situation

[Anonymous Coward]

For Linux to truly become the gaming OS of choice it will need a killer app that can't run in Windows, forcing users to switch over.

Problem is, no developer will be willing to develop said killer app until Linux becomes the gaming OS of choice.

Games? Yes. Games distro? Who cares?

[phrenq]

Windows became the top gaming platform without any special "gaming" versions of its OS. They did this through marketing and its DirectX APIs. Get some good games and people will play them regardless of their distro. Get a "game" distro and nobody will use it without good games. Either way, the distro doesn't matter.

A few ideas to throw out there...

[FyRE666]

It would be quite cool to have some game-targeted features in the kernel for instance:

Ability to "lock" the scheduler, so that the game gets 100% CPU until it unlocks (effectively
making it a single process OS like DOS while in this mode).

While in the above mode, a user-configurable keypress to pause the whole system, no matter
what's going on.

Running the games in kernel space? Maybe this is just madness ;-) Would it not help performance
if the CPU wasn't switching between contexts?

I'm sure I could think of more - yes I know this might not make the most stable system out
there, but for games use, wouldn't that be a good compromise?

Re:A few ideas to throw out there...

[ZorbaTHut]

Windows actually had this for a while - DirectDraw5 or so recommended that you lock the Win16 mutex when locking the back buffer so you could return it as quickly as possible.

They don't have it anymore. It provided, at most, maybe about 0.1% CPU boost, and if anything went wrong during that time *boom* the entire system melted down. I imagine, yes, they could have fixed that, but it would have been buggy code.

Are you seriously telling me that your Linux box runs at a load factor of more than 0.01 when it's not actually doing anything?

20:24:40 up 5 days, 8:58, 1 user, load average: 0.00, 0.00, 0.00

I don't see the point of increasing complexity, or moving things into kernelspace, just for an extra 0% performance. :P

Just perfect Wine/WineX

[mao che minh]

Realistically, gaming companies probably won't be writing many Linux games for a long time. Sure, there have been a few games here and there (Unreal Tournamemnt, Wolfenstien, Quake3), but they usually launched long after their Windows counterparts' releases. There is a good reason for this, considering Windows monopoly and an almost non-existant Linux gaming community.

I think the answer then lies within a solid emulator. I think gamining companies would support this as well. It would take them far less time and money to make sure their game was programmed to operate within Wine than to write a Linux port. Not to mention the pool of open source volunteers at their disposal.

Re:Just perfect Wine/WineX

[Minna Kirai]

Even though "perfecting" WineX is an enormously challenging task, the problem is not even that simple.

The most important new PC games coming out are multiplayer online games, and they're starting to standardize on Evenbalance's Punkbuster library as the way to prevent cheaters from hacking their local environments with transparent walls and magic maps.

Punkbuster works by examining the entire memory environment where the game is running. If it detects something that could be a cheat attempt, it shuts you down (optionally blacklisting you with the publisher's master server). It's constantly updated to respond to new threats.

What this means is that game publishers soon will not want you to run under Wine, and will pay programmers to ensure that you don't. (For example, Battlefield 1942 used to work in WineX. Since Punkbuster was added to the game, it's no longer usable)

To prevent cheaters, game makers have decided to allow playing only under a "trusted" environment. They won't allow you to play from an Open Source OS or emulator, because that opens up the possibility that you've changed the graphics driver to make wireframes instead of solid textures.

Re:x-box run linux?

[Quarters]

You're right and you're wrong. You're right in that the XBox does not run Linux as its core OS. Geeks have hacked it on there, but the box does not natively boot into Linux.

You're wrong in that the XBox runs a highly modified version of the Windows2000, not CE kernel. Just enough of the OS for booting, hardware configuration (aka Live! config), and DVD autoloading is kept internal. The rest of the libraries required to run a game are loaded off of the game DVD.

If you read, "Inside the XBox" you'll know that the original spec was for a custom version of WinCE to be used, but that was scrapped since it would've required making a fork of DirectX that worked with CE.

Re:Bootable CD is the answer.

[Anonymous Coward]

gentoo already has bootable game cds, one with americas army, and another with ut2003 demo

A cross platform game would help

[Cap'nMike]

Imagine if the open source comunity were able to develop a couple of really good games, say just an FPS and an RTS, then release both windows and linux versions. The catch being to charge for the windows version, while releasing the linux version for free. If the games are good enough and don't focus on the activities of penguins, this would be incentive for windows gamers to try linux and see the benefits. I know that the games would then not be considered "free", but the developers could still release the game engines under the GPL or whatever.

Better yet - make it a multi-media console

[falser]

A modded Xbox running XBMC is a whole lot more user-friendly than anything I've seen for Linux. The software is easy to configure and use, looks great on an HDTV. As I understand XBMC is a port of mplayer - but the customizations they've done for it to work with the remote control and adding a multi-media browser (for file selection) take it to the next level.

What would be really great is to port XBMC back to Linux, and meld it with MythTV for PVR functions. Supply the distro with preconfigured Emulators (just drop roms in a particular folder). I'm sure a distro like this would be something that many people would be interested it.

A text adventure would be perfect

[WankersRevenge]

Just think about it. You boot up Linux for the first time, and the way to activate functionality is to make your way through the "game". The first thing it should read when you boot it up:

It is dark. You will mostly likely be eaten by a Stallman.
>inventory
You are carrying:
man light
>man room
The room brightens. You are in a small chamber. A sign on the wall declares this room to be: init.

A door reads, "Daemon Restroom". A light glows from underneath it. You hear a toilet flush

A tall lanky fella steps out of the darkness. He wears a threadbare cloak and carries a large sack. He opens the sack, and grumbles something about "699". A large stilletto knife dangles from his belt.

I believe...

[SwansonMarpalum]

Ideally what Linux needs to do for game developers is offer them something more than what Windows gives them.

What could this possibly be? Imagine putting a game you just bought in your computer and it booting up with an OS which is minimalistic with regards to the game in question. Everything it needs and nothing more. Whatever overhead there might be in Windows is irrelevant, this OS is there and just does exactly what you as a game developer needs.

The system boots from the CD (ie knoppix), mounts your windows Hard Disk read/write for game saving, and loads the game. If it's a network game, it brings up your network interfaces too. Everything is detected, and the OS is configured the way the game needs it.

TO BOOT (no pun intended), you can also install the game as a normal windows game and run it from the windows environment if that's what you want, as a user.

Where could one obtain an operating system where they could build this bootable CD from and redistribute free of licensing fees??

What the OSS community who is interested should be focusing on is providing this technology for game developers, giving them a clean and robust migration path out of Windows. Then, miraculously, this framework can be put on top of your existing Linux install with no effort.

Call me crazy. ;)

EverQuest

[Fizzl]

Just port the f*cuking EverCrack onto Linux and I'm ready to migrate my desktop.
No seriously, that's the only thing that is keeping me and my wife back. :)

Re:x-box run linux?

[polyp2000]

I think that the poster is obviously refering to the X-Box linux project, which via a buffer overflow exploit in certain games, enables linux to be installed without requiring a mod-chip.

The inherent open-ness of Linux and its various development kits allows developers to create software and games software without the costly restrictions and control console manufacturers seem to place over their respective hardware.

It should be noted however that Sony have released a Linux based distro specifically for Playstation 2 for exactly this purpose.

My own personal belief is that it is extremely difficult to create next-gen games without the kinds of near-hollywood budget software houses have to throw at it. Im not saying its impossible, but small scale bedroom coding aint gonna produce the kinds of masterpieces that Lionhead or $GAMESTUDIO_OFCHOICE are producing.

I think a better twist on this idea would be to produce bootable CDROM's ala knoppix, bundled with a specific game. This way you remove the notion of operating system dependancy. Linux enables you to build a very low-level OS, with just enough required to boot the game. If something along these lines were to be introduced it would allow mainstream software studios to sell games to anyone who has an x86 machine, regardless of OS.

Comment removed

[account_deleted]

Comment removed based on user account deletion

Chicken V Egg, Round 300

[inkless1]

Games don't attract people to an OS - an OS attracts game developers because of a target audience.

If Halo had come out for only Linux, do you think there'd be a million more Linux users? No, because nobody is going to ditch their OS just to try out one game. And no game developer is going to spend the millions it takes to make a AAA game on an OS with low yield.

Maybe, just maybe, if there was an excellent hobbyist community and development platform then as amatuer productions like FPS mods and the like get more and more mainstream Linux could get a bit of rise up, but nothing serious I'd imagine.

Linux should just keep the long slow road it's been on. Get prettier, get friendlier.

wait..

[CoolMoDee]

wait..did you just say XFree86 and developing fast in the same sentence? Best laugh I have had all day thanks :-)

Obviously

[foonf]

Its common knowledge that becoming a superior gaming platform is the best way for a platform to gain mainstream acceptance. Thats why the Amiga has become the dominant computing platform today.

Re:Multiple monitor support? Has been there.

[whodunnit]

First off,

Please do some actual research before you state something as a fact. An accrual informed write-up of multi Monitor support in windows

And that review focuses solely on gaming under multi monitor situations in windows, there are even more options available if you are not trying to game. So your "only 2" options in windows statement is quite false. Thanks for your time.

Everyone is root?

[gotr00t]

The kernel can definately be hacked so that it allows this, but this presents a huge security concern. Every user would have to have the same priveledges as root in order to do this, or the user must play as root.

Yeah, whatever!

[Otter]

If in the FOSS community we could only get our act together and launch a game-based distro, we will be home and dry.

If the people who run around holding forth on what "we" need to do actually did a tenth of what they're calling for, we'd be "home and dry". For that matter, if they did anything useful, it would make all the difference.

Honestly, we've been hearing "What we need to do is make the bestest game ever and only sell it for Lunix and then everyone will use Linux!!!" for years. And what "we" have to show for it is Tuxracer and 500 libraries in search of developers.

What's missing from Linux games? ORIGINALITY!

[Junks Jerzey]

Linux games have always had a very me-too nature. There are emulators for old systems, old commercial games that have had the source code made available, lots of little hobbyist remakes of Tron Light Cycles and Boulder Dash and some C64 games. There's some other stuff, too, but not much.

Back when the Apple IIgs was dying, and I paid attention to that system, there was a similar pattern. Oh so many programmers wanted to prove that the gs was an awesome system, so what did they do? They wrote clones of games that were available for other systems. Really, this was cool for the people who only owned a gs, because they couldn't play those games otherwise. But as an outsider looking in you saw all these versions of Tetris and Lunar Lander and so on. Some were spiffy, yes, but wow did it make the gs seem stale. The Amiga followed the same road. It would have been much better for the programmers of those systems to lean hard on creativity rather than getting in a pissing contest with other computers.

One Word

[Meneudo]

Pippin.

Remember the short lived Apple console? That's what happens when a company without many resources tries to enter the game industry. Although this isn't a console approach, I doubt it would end in success. People won't flock over to Linux just to play games. Nobody ever buys a Mac for gaming.

Linux already has a market niche and is associated with being 'for nerds.' It's going to take a serious overhaul to try to do this, and its not even guaranteed to succeed.

Linux games myth...

[msimm]

I've been using Linux as my sole home desktop environment for years now. Since the very begining we have been hearing (and chanting) claims about how Linux needs game to become mainstream. Whats interesting is Linux now *has* games. I think a games focused distro would be smart, but certainly won't fix (or hide) the number of other areas in which Linux distro still need to mature.

Linux isn't experiencing a high rate of adoption because its still too hard to use. We know this. No amount of games is going to fix that and [name your favorite distro here] are making slow but relentlessly steady headway (see Microsoft cringe).

My point is there is no single solution at this point. Linux needs Users Friendly standards from the layout to the message dialogs, application naming conventions, install/uninstall and system configuration. Thats a lot detail and involves a lot of seperate pieces. Standardising is also FUCKING BORING WORK. So don't expect it to happen as quickly as some other things.

Games are cool, but its not that simple.

Games On Bootable CD or DVD

[MS_leases_my_soul]

I have been asking this one since the day I saw my first Knoppix CD.

Why can't we build games where everything you need to run the game is right on the CD?

There are already Linux distros out there that boot into MAME. Why can't we create some type of standard that is the "whole package" answer to DirectX?

As long as your hardware is compatible, you just work. You boot from the CD and play that game and that game only. We can create a standard bootable game distro and port games inside that distro.

Once you have it running in a "fixed environment" of a bootable CD (you know every piece of code on the CD and its version, so you are in total control of compatibility and run environment), you can expand to get the same game to run in a general Linux environment.

Would it be a PITA to reboot my PC just to play a game? Yeah. Don't I already do something similar with console games? Yeah. Aren't I basically just turning my PC into a fixed environment like a console? Yes, but it is an environment where the developer has total control over the run environmnet.

Am I smoking crack here or does this make at least some sense?

Okay, if that's what you want...

[Chris Burke]

There should be a unified development tool/library that includes them all. E.g. I can install "blah" and boom I got 3d graphics, sound support, joystick/keyboard support, timers/interrupt/callback etc...

Okay then... I'll just take OpenGL, SDL, and ALSA, put them in one Debian meta-package, call it Universal Games API or "blah" or whatever makes you happy, and there you go.

SDL, OpenGL, ALSA all solve one problem well. They also work together well. Writing OpenGL apps using SDL is simple.

I'm not really sure what you want or why you want it. Yes, all of these libraries are "UNIX philosophy". That means that not only do they do one thing well, they are designed to be easy to make work with other programs that do other things, so you can easily get one program that does both.

What more do you want?

Re:Why I don't run Linux

[dr bacardi]

I'm gonna agree with you, mostly, but (and there has to be a but ;) I don't think that you just outa the blue one day decided, "Hey, I'm gonna edit movies!"

I'm certain that there was a learning curve, and for some of the higher end NLE's, it's pretty damn steep, and for me, "too much work." But then again, I've used Linux for a long damn time, and *to me* it's actually easier to set up a wireless card under Linux than a Mac(*). So, in the end, what "too much work" means is different for us; just because you are used to one way doesn't mean that different == useless.

And if you can afford 10k$ + 1.5k$ a year for Shake, give the Linux geek down the street a couple of hundred bucks to set up your cluster ;)

(*) Granted it was a Proxim Skyline w/ crappy drivers.

Why limit ourselves to just GNU/linux?

[trs-sld]

Why doesnt the OSS community collaborate with Apple to make a robust *well marketed* alternative to DirectX for *nix? It would use OpenGL of course for the graphics. The rest of it might even be able to come directly from some existing projects.

This would be a win for Apple and the community as then game developers could target one platform that would encompass Mac, Linux, BSD etc. Perhaps the combination of all these platforms together would be a big enough number to start convincing game companies to pursue the *nix market.

The key here would be convincing Apple to throw in the marketing. Without marketing, it would probably never take off. And come to think of it, maybe it would be impossible to convince Apple since they really arent trying to sell gaming machines. idunno, just a thought that seems to make a lot of sense in a lot of ways.

I Disagree

[Greyfox]

The biggest challenges are to get automated hardware detection and configuration working reliably and easily.

Friend of mine recently decided he wanted to fiddle 'round with Linux. First thing he tried to do was Debian. After futzing around trying to get X working for about a week he gave up and wandered off.

He came back with RedHat 9 which did do a pretty good job of getting X working, but it was godawful slow. I suspected he needed the latest nvidia driver off their web site. He wandered off to get that, then wandered into the twisty maze of package dependencies he needed to get it working. RedHat could take some pointers from Debian in the package dependency arena (That's why I kicked them to the curb last time I used the distro.)

My friend wanted to be able to play assorted video in Linux too. Pretty sensible. So he started looking into mplayer. Now, I know there are a lot of legal issues surrounding mplayer, but it's kind of difficult to explain those issues to someone who's used to just having the ability to do all that stuff in Windows. He wants to just install the package and have it work. He doesn't want to have to locate DLLs in 18 different countries and compile code that may technically be illegal here in the States to get it working.

So there's step 1. If I can slap a Linux bootable CD into pretty much any system and have it boot reliably, detect all my hardware reliably, and provide accelerated 3D and play video without me having to compile a kernel I will consider step 1 a success.

Step 2 is providing the libraries necessary to write the software for Linux. Look at all the major consoles and Windows itself and what do you see? Those corporations sell a SDK to people who want to develop software on their platform. Do the software libraries that are available for Linux compare favorably to the ones for the other platforms? I'm pretty happy with the Linux application libraries, but games have specialized needs.

If you provide those two things, you've got the beginnings of a cross-platform gaming environment that a lot of gaming companies should find very interesting.

Game disk

[Christ-on-a-bike]

On a console, you put in the disk and your system boots straight into the game. Why not have such game disks for PC systems? The (stupid) reason is that Windows can't be distributed cheaply enough, and everyone writes their games for Windows.

Linux is free. It can be included on a bootable disk with your game. So while hardware remains an unkown, at least your game can run on a known kernel, known libraries, optimised X server etc. Swap space (if needed) can be automatically found in Linux partitions or Windows swap files.

Managing players' saved data is the biggest problem here. A nice solution might be to save it over the internet to central servers. Now they can load their saved games from anywhere, and play on any PC.

Of course the hardware detection would have to work more flawlessly than Knoppix, not an easy task. This method of distribution would not suit all games.

I can beat that: 2000+ games on a single Linux ISO

[leonbrooks]

It's called AdvanceCD, plus you need to haunt the alt.binaries.emulators.mame usenet group for a while (try aub and see if your ISP carries it).

Also, the MS page linked above is for their optional "Plus" pack, not for the base XP system (which comes with, what? Solitaire, hearts, minesweeper? Do we now have a more advanced MCSEHS qualification? - Minesweeper Consultant, Solitaire Expert and Hearts Shark). I do notice an ominous counter to one FOSS advantage, though, a "365 tips from users like you" section.

Static is not a good idea

[lysium]

Why the hell don't people staticly link this stuff? The APIs for many libraries are so unstable that the idea of "What if I wanna update libBlah later on?" doesn't work and it's not all that important that save on transit or hard drive space any more.

Another aspect to consider is system security. If every app on a linux system came with static libraries, then you have multiple libraries scattered all over the drive. Will all those application authors update their program to include library updates? What if a nasty buffer overflow turns up in libBlah...do you want to leave all the dependent programs around for crackers to stumble upon?

I am not saying that the convenience factor is not important; rather I think that an altogether different approach is needed, one that tackles the problem at a different level. Development on ports systems (Gentoo) is one interesting direction, autopackage another. Better that than applying static libraries to a problem they were never designed to fix.

===---===

Static *and* dynamic is best

[Shinzaburo]

One of the biggest problems with *nix systems are dependencies. This is a problem that would go away if all applications were distributed as self-contained packages, a practice that should be the default behavior when distributing software applications. With few exceptions, anything that requires the end user to download pre-requisite software when it could be easily bundled is, quite honestly, just plain silly.

"What about security? What about performance?"

The app should be designed to give the end user a choice: Do you want to use a dynamicly linked library? Fine -- tell us where it's located and we'll ignore the stuff we thoughtfully bundled for you. Do you just want the damn thing to work? Yes? Fine -- you don't need to do anything further, and we'll just use the bundled libraries.

"What about disk space?"

Given the benefits of software that just works, a few extra MBs of space is not even worth wasting brain cycles on. For those that feel otherwise, I suggest they figure out a way for apps to be packaged such that undesired bundled libraries could be easily jettisoned.

This isn't La-La Land that we're talking about here -- just look at Mac OS X. Most applications there aren't even "installed" in the *nix/Windows sense of the word; the end user downloads the package and drags the application icon into the Applications folder. Done. Any dependencies are contained within the .app bundle. This is the way all software should work.

If application developers would all agree to do this, the world would be a much better place.

Simple Direct Media layer

[Jonner]

SDL is what you're looking for; it's been around for several years. It's mature and in use in many projects. I don't know everything that DirectX does, but I believe most of it can be handled by SDL combined with OpenGL. Not only does SDL run on many platforms (including Windoze), it has bindings for various high level languages, so one isn't stuck with C or C++.