Tweekers' Blog

Many view gaming as the biggest obstacle keeping them from removing Windows from their computers. I can 't count the number of times I've things such as “I keep a dual boot because I still want to play some games on Windows”.

I myself like gaming but the reasons I don't use Windows are 1) that I'm not into the latest games as the majority of PC users are and 2) I'm willing to hack for hours in order to get a game working (what regular users want to avoid at all cost). But to be honest, I'm not a hardcore gamer, and that must be the reason that I've easily given up on Windows.

When a user gets into Linux there's not that many games to choose from (I mean big commercial titles working out of the box). Sure there's Open Source games, I'll talk about them later, and sure there is Wine but this just isn't a reliable solution. Take for example Silent Hill 2 which is a really great game by the way, it works flawlessly in Wine and I had a really good experience playing it in Linux. Silent Hill 3 is also a great game, and technically it's not that different from the second episode so one would say it should also work in Linux. Well no,  it just crashes (at least it did when I tried it, it seems that there is better support for this game now) so in order to play this game I had no choice but to play on a Windows computer. And that's what the majority of users will do, regardless of the game they want is rated Platinum or rated Garbage on Wine's AppDB.

Speaking of these ratings, these are just unreliable for any normal user. Many Wine testers have really low expectations and they might think that rating a game as “Gold”makes Wine more prestigious, but really it doesn't. When you play a game rated Gold, you expect some good results, but sometimes all you get is a framerate 2 or 3 times less than the one you get in Windows. Of course this rating can be legitimate sometimes, some users submit tests result that have been done with a big fat Core2 processor and a big fat Nvidia G2xx but totally forget to mention it in the results, and it's kinda hard to see the difference between 70FPS and 200FPS : both are just really playable.

There are also many games with a good rating where the tester tells us the game runs pretty well in 640x480. Come on, it's 2009, I would think no one judges playing games in such a low resolution as an acceptable solution but sadly, many do. Same thing goes with the level of details the game let you use. If you turn off HDR and put low resolution textures, it's not how the game was meant to be played.

It seems that these kind of tests do not even comply with Wine's rating guidelines. I remember that in the past, ratings where based solely on compatibility and not performance but now, reading from the current guidelines, an application running much slower than expected should be rated Bronze.

This doesn't change the fact that Wine developers target  better compatibility and not necessarily better performance and that's where Cedega got my attention. Cedega is the closed source version of Wine and aims to provide a better support for 3D games. For many games it does the job pretty well, the best example I have in mind being Left4dead which is unplayable in Wine but totally playable in Cedega. Cedega might be non free in the both ways, it brings gaming to the next level, providing support for many recent games.

I've been talking a lot about Windows games but nothing gets as good as native games and when I think about the current situation all I can do is cry in despair. The only recent titles that I can think of with a linux port are Quake Wars, Prey and World of Goo (and Prey is already kinda old). Other than this, there is Linux Game Publishing that sells ports of old games that I've never heard of before (Sacred is from 2004 and X3: Reunion from 2005). I miss the old Loki days in 2000-2002 when they ported really good games like Soldier of Fortune, Quake 3 or Unreal Tournament. Now all is left of Loki is Ryan 'Icculus' Gordon how seems the only guy on the planet really capable to do a Linux port.

The major question I asked myself is why is there less native games now than in 2002 ? Well, I have some clues about this. The game market has evolved since then and it has become much more expensive to make a game. The days where games where made by a small group of hackers are over and the current gaming industry is now even bigger than the movie industry. Game editors want to be paid for their games,  and the Linux market is just too low to even think about a port. I've been reading the Linux support thread on Unity3D forums and it's really enlightening to see what professional game developers think about the Linux community.

First major problem : the marketshare is way too low to even think about developing on Linux. The real problem is that this marketshare is low because Linux has almost no games to offer so the dog bites its tail here and we're stuck in an infinite loop. Arguments like “There is no DirectX on Linux” or “Linux is missing this library” are totally invalid. Game developers can do ports for any platform when they want to. Saying that Half Life ² will never get to Linux because it uses DirectX is wrong, HL² has been ported to the Playstation 3 and I've never heard of DirectX on Sony's console. When you look at the Sega Lindbergh arcade board, you see it's nothing more than a PC running Linux, still we don't get to play The House of the Dead 4 or Virtua Figher 5. All the latest generation consoles run on IBM Cell architecture and most of them don't use DirectX, meaning that every game is portable if you really want to.

Another problem adds to the problem of having a 1 or 2% market share of the whole desktop PC market, it's that Linux user have a reputation of not paying for stuff (because most of their software are free as in free speech AND free as in free beer). I think that is mostly untrue, at least in my case. I bought a few games on the Steam platform but how will Valve be aware that I don't run their games on Windows ? I think this reputation comes from the fact that you hear a lot from Free Software activists on the Internet whereas casual users, who are potential buyers, tend to stay quiet.

A user on the Unity 3D thread sums it up like this :

“Linux (desktop) users are (a) smaller in number even than Mac users, (b) buy less software than anyone, (c) have ridiculously high support costs unless they support themselves ... feel like GPLing your codebase? ”

The last point this user made is also a big problem. Sure there are some application that you can just download from a website and they will work on any distribution : think about Firefox, Google Earth , Eclipse or Doom 3. But as the application ages, it gets much worse when we're dealing with closed source software. Think about those old Loki installers, it's horrible to get them working on a recent Linux distribution : you'll have to deal with OSS, glibc 2.1 compatibility, and these sort of obsolete stuff. Quake 3 is a good example for this problem. When running the official closed source installer from iD software I'm pretty sure you'll run into some kind of problem at some point. If you use the open source ioquake3 binaries then you can except it to run out of the box. So yes, GPLing your codebase is indeed a good idea. At least game developers  should do it when stopping to support their applications.

I don't think that any game company has brought more to Linux gaming than iD software, they have written great game engines and more important Open Sourced them after a few years. Even though, on the same Unity3D thread I can read

“John Carmack of id Software has been quoted as saying that porting to Linux isn't worth it. Even though id does it, they do it as as technical exercise and not because it makes them any actual money.”

and

“ They started out developing on NeXT machines (no kidding) because they understand that you develop on the machine you like and ship on the machine you have to. They make so much money (and have so much technical skill) they can afford to take that attitude. Most games companies fail in both departments.”.

This is far from encouraging, knowing that the next engine id tech 5 might not be ported to Linux. Linux marketshare is slowly growing but game companies and offering less support than ever ! We already lost Epic Games with their failure at providing a UT3 client for Linux (while I think it's more Midway's fault than Epic's), if we also loose ID Software then ours chances of seeing future games on Linux are close to zero.

So what's left ? Open Source games ? Well excuse me but I'm a bit skeptical about these. Two years ago I thought that Open Source game where technically inferior to their commercial counterparts. Engines like Cube 2 (used in Sauerbraten) and Darkplaces (used in Nexuiz) have proved recently that they where capable of great things while others like Crystal Space / Blender Game Engine used in Yo Frankie! didn't impress me much, giving me the impression of an engine dating from 5 years ago. But I'm okay with poor graphics, I enjoy a lot of old games that don't look nearly as good as todays games. The only thing I ask for is good design and good gameplay or in one word: immersion. This is where Open Source games fail, I've never encountered a Free game that I wanted to play from start to finish if it's a single player game, or challenge me enough to keep me playing if it's multiplayer. I can't figure out how a large community of contributors fail at providing good design and gameplay. I guess that talented game designers get in the game industry and get paid making commercial games.  One good example of this Open Source failure is Open Arena. I've played Quake 3 Arena for hours, finished every level in Hardcore mode, finished some in Nightmare mode, I find the maps fun to play in, the models look cool, the sound effects are memorable. When playing the Open Source remake Open Arena is always got bored, the maps are awkward, the models and sound effects are pretty lame and more important than all, I don't have this feeling of playing a game with the perfect gameplay (Quake3 was pretty awesome for that). Actually, the best hope in Open Source gaming is Nexuiz, the gameplay is almost good (they still have to work on the movements and the part dealing with taking damage), and while I find the models poorly designed, the gameplay is fast enough so that I don't care about them that much. At some point I though that I was just beginning to get bored with video games in general. But then I found great new games that I enjoyed : Portal, Team Fortress 2 and Left4dead. That's what I call great design, perfect gameplay and total fun !  The new Prince Of Persia was really cool too and I enjoyed playing this great platformer.

Looking at the forums and blogs, I can clearly see that there are gamers on Linux that actually like all these Open Source games. These people are mainly the ones who reject all proprietary software. And while I understand the dislike of  proprietary code, I don't get why some people also insist on having Free (as in speech) content. Even Free Software activist Richard Stallman doesn't care about Free artwork ! Some activists gamers got carried too far away, I just think that the only reason they want Free games is to get free (as in beer) games. Free Software is not about this, it's about having the control on your programs, and we cans see why game developers deserve to get paid.

I think that games with proprietary artwork and gameplay running with an Open Source engine is the only hope left for Linux gaming. Imagine that Valve open sourced their Source Engine (while they could start with their HL1 engine …), would they sell less Orange Boxes and Left4deads ? I don't think so …We would see Source games ported to Mac and Linux in a few months or weeks and then Valve could see that making an engine Free Software does increase sales. Of course this method doesn't work for companies that only sell a game engine to other game developers like Unity3D does. In these cases it would be a good idea to use iD software's method and Free only the engines where you don't make anymore profit. Of course, in order to deliver a Free game engine, a game company is required to care about it's costumers, and we all know that's pretty rare : They're mostly here for the money, we're not in 1999 anymore, back when games were still made by a small team of passionate hackers. Open Source games also raise the issue of not having malicious code like DRM, in the game. I don't think game published like the idea of  giving their costumers this amount of power, neither do they have the desire to bring unlimited support for their games (Open Source code can be maintained forever), because supporting old games can mean selling less of the newer ones. Given all these facts, Linux gaming is in a pretty bad situation today, and maybe that video games aren't compatible with the spirit surrounding Free Software and open systems like Linux. In the book “The cathedral and the bazaar”, the author Eric Raymond states that 90% of all software development in for in-house use only, and using Free Software as proved to be the  most efficient method here. Sadly, video games are part of the remaining 10% and while I think that most the value resides in gameplay and design (which have no point in being Free), I'm not sure that an editor would be willing to Free its codebase just to gain a few Mac and Linux users.