Ryan Gordon's Linux Game Porting Guidance

[phoronix]

Phoronix: Ryan Gordon's Linux Game Porting Guidance

Well known Linux game developer and porter Ryan "Icculus" Gordon spoke at this week's first-ever Steam Dev Days about porting games from Windows to Linux. For those curious about what's involved in porting a modern game from Windows to Linux, he's shared his slides to cover the process...

http://www.phoronix.com/vr.php?view=MTU3Mzc

[mrugiero]

Nice article, Michael. I'm gonna be reading the slides now.

[FLHerne]

i

• QtCreator is the best MSVC alternative IDE out there

No, no, KDevelop is!

[peppercats]

Just a few more alternatives:

• QtCreator is the best MSVC alternative IDE out there
• GamePlay3D and NeoAxis are good alternative game engines
• GLFW is a good SDL alternative if you are looking for a more light weight library

I don't think neoaxis has any linux support?
Leadwerks is really good, but it's not out for linux yet(should be really soon.)

[IanS]

Leadwerks is really good, but it's not out for linux yet(should be really soon.)

I'm not seeing it, it seems fairly run of the mill to me, from what little I can find on it. To make matters worse it has a pretty shitty license that prevents you from making your game modifiable by anything other than their own software, which means if you want to make your game a bit more extensible for users to bring more life to it, it can only benefit those who fork out the cash for their own Leadwerks license, which is pretty lame.

With any luck though that Godot engine that is to be released publicly next month should make Leadwerks irrelevant. It seems to be an older and fair bit more mature engine, plus it will be free and open sourced under a very liberal license (MIT).

[wargames]

Ryan's stated advantages in bringing your game to Linux is that the ecosystem isn't "flooded with cheap product, like Windows. You don?t have to compete to be heard over the mob." Additionally, there's no dictator controlling about what can be published nor taking a cut of sales -- like app stores for other platforms. Last and perhaps most importantly now, SteamOS is Linux based.

What? Some dictatorship is actually good for everyone. Look at the Android apps, most of them are pure junk/spyware/ad-crap(TM)

[IanS]

It isn't suitable for serious PC gaming (yet), because it lacks deferred shading/rendering backend. Their current method (forward rendering) suits devices with low video memory.

I find deferred lighting/rendering to be a bit over rated most of the time and it unnecessarily cuts out a large portion of the hardware out in the wild from playing your games when in most cases proper use of light maps will get you perfectly fine results for a fraction of the overhead. None of the game projects I have in mind for the next few years will be making much (if any) use of dynamic lights so I am not all that fussed that they haven't added support for deferred rendering just yet.

[peppercats]

What? Some dictatorship is actually good for everyone. Look at the Android apps, most of them are pure junk/spyware/ad-crap(TM)

Easily dealt with by user reviews, most ad-crap has horrible reviews.

[IanS]

...

I understand how forward and deferred rendering work and the main pros and cons of each. Deferred rendering is mainly of use in games where fairly heavy use of dynamic lighting is desired. So ya, I suppose if you are mainly interested in FPS games with realistic dynamic lighting then deferred rendering is much more important, and if that is the case then Leadwerks is good enough (as long as you don't mind the limitation on modding). It would almost seem more appropriate to call it FPSMaker, as it is to FPS genre what RPGMaker is to RPGs. FPS games though are a dime a dozen, and I am not all that interested in playing any more shoddy ones created by amateurs which probably deepens my bias against it.

At the moment I am primarily interested in puzzle and 2D/2.5D game ideas and I am partial to an NPR art style. Most of the games I plan to make in the next few years will mainly be be done with pre-rendered 3d content and fixed camera perspectives. For the most part any lighting needed will be baked in and the only things that might need real time lighting will be the player/npcs/mobs (depending on the art style of a given game). So deferred rendering is just not of much use or interest to me at this time. Also because of the sort of games I am interested in working on, Godot has a lot more potential to be a good fit for me while Leadwerks wouldn't even be an option due to its FPS focus.

[stqn]

Good stuff. It?s nice that Ryan commented the slides (although the text was really small), because often conference slides contain too little information to be useful.