Comment by sandGorgon | Hacker News Reader

Comment by sandGorgon | Hacker News Reader

Parent Full thread

yes it will - Steam is a juggernaut which will trigger purchase/usage decisions. Creators of libraries and frameworks will have to comply with the standards that Valve will come up with (and come up they will).

For example what audio framework will your distro use - Jack? OSS? Pulseaudio? They will have to use what Valve wants you to use, if you want to play their games. Multimonitor support ? HDMI? use the toolkits that Valve will define. I would'nt be surprised if Valve influences the switch to Wayland or something.

And the big one ? packaging format.

What I'm really hoping for - Valve being able to arm twist ATI and nVidia for better drivers (which may happen anyway because of Ivy Bridge graphics and its great linux support[1] )

This is a Good Thing. Linux has been suffering from a severe framework fragmentation which leads to a very small addressable market. Unifying behind some frameworks will undoubtedly increase that market size.

[1] www.phoronix.com/vr.php?view=17284

No it won't. Login to steam from a mac some time and watch the awesomeness unfold as 1/10th of your games are available.

Now make it 1/10th of 1/10th!

Unless they convince a significant number of other publishers to do linux versions (while they won't even do mac versions) it's not going to be very exciting.

What if Valve expanded "Source" from an engine to more of a platform? If you create your game for Source you can release on Win, Mac, Linux, and a SteamBox/console with relative ease?

It's pie-in-the-sky, but that would be a game changer. A linux steam client would be one piece of that plan.

Yes, a lot of people throughout the world are working on write once run everywhere and they're the Java, Go, Python, Ruby, Lua... language developers.

And I have to imagine that what you're proposing is Valve write some virtual Steam-Machine that they port to every major OS, then they write a compiler that outputs Steam-Machine byte code. And maybe that could work (though different graphics libraries seem like a very hard problem to overcome). And maybe it would work really well for smaller indie games like Super Meat Boy, Braid, etc.

But I don't think it will work for AAA, computation intensive games life LFD2 or Portal 2. I don't know much about game development but from what I've heard, if you're making a game with whoa-mygod graphics then you're doing some really close-to-the-metal programming trying to save every cycle you can. And all the effort (and results) are for naught if you're running your super optimized code on a virtual machine. Portability would be amazing, but are people willing to cooperate and say "I'll live with a game that looks last generation because so someone running a different OS than me can also enjoy it."

I don't think so.

Disclaimer: I'm not a game developer, I could be wrong about the technical challenges, and if so, please correct me.

Much as a game built on Unreal or Frostbite can trivially support Windows and a variety of consoles, a game built on a future version of Source could easily support OSX and Linux.

Platform-specific code tends to be in the guts of the engine, not the content and gameplay code that characterizes a particular game running on it.

I'm not proposing a virtual machine or a new language; C++ and the Source engine seems to be working fine for LFD2 and Portal 2 and other AAA titles.

I'm proposing liberal licensing for the engine and adding a simplified API for 2D games. You are correct that you can't optimize for every platform at once, but that's what #ifdef is for. Extra work is still required on each platform... if you need the extra performance.

Mostly I can't imagine Valve putting the effort into Linux if they don't have a plan for getting more games onto Linux. I was just trying to guess at that plan.

Audio: ALSA would cover basically all distributions. Pulseaudio is used or available for all recent releases. Plus, OSS, ALSA and Pulseaudio aren't exclusive. You can have all of them available at the same time.

Multimonitor support and HDMI have absolutely nothing to have with applications.

Packing format (for Steam) is irrelevant when you're just distributing statically compiled binaries. And the games themselves would be packaged in Valve's DRM-enabled format anyway, like on the other platforms.

Doesn't ALSA have some pretty big limitations?

ALSA as the only thing running on your system does. But if your game supports ALSA then it can plug into ALSA which plugs into PulseAudio and you still get all the benefits of PulseAudio.

I always uninstall pulseaudio as the first thing I do on my Linux box.

I have always noticed a latency between the time that a sound is supposed to be played and when it is actually played with pulse. This doesn't matter so much when listening to an MP3 or whatever but if you are trying to use a sequencer or play a game it becomes a deal breaker.

What are the benefits of PulseAudio other than making your sound system buggy and increasing audio latency?

Individual volume control for each app. That's why I use PulseAudio despite its problems.

dmix?

You can connect to a remote audio server. I've used this feature to output sound from my laptop through my desk speakers even when my laptop is unplugged from its docking station.

its not about irrelevant, its about changing mindsets and acceptance. If Steam can work out of statically compiled binaries, why not use it throughout ? (that's not a question, but a thought) Multimonitor support is VERY,VERY important for a gaming application. So is HDMI. Some games autoscale when they detect a HDMI connection - will you particular distro be able to do that ? If not, then I will switch to a distro that will.

The point is that there is nothing implicitly wrong with Linux as a technology - there is something lacking in individual distros who seem more fixated on IM frameworks than networking/sound/audio/peripherals. For Steam, these are paramount - and this will influence distros.

How is all this going to happen? Steam is just the thing that downloads a game and runs it, it has no other say in what your game is written with or how it runs or where.

Valve/Steam is not a SDK. It's just a sort of package manager. What you say makes no sense to me.

As long as Valve wouldn't be pushing closed/proprietary standards or other nonsense, I would be very okay with this.

What you're talking about is the Linux desktop world and you could be right about that; maybe it will finally become more standardized but I don't see mainstream desktop markets really noticing. Linux desktop is, what, maybe 1.5% or so? This won't be the magic ingredient that suddenly makes them take over the desktop.

No, but a jump from 1.5% to say 4% could be more significant than you think, especially if these are the right kinds of users.

Remember that a significant number of Windows users are either corporate users and corporate use numbers could take a significant fall if 3 or 4 big corporations decided to switch away from it. Many of the other windows users are people who bought they cheapest PC/Laptop they could get and spend as little as possible on software for it.

If Desktop Linux had 4% but that included a significant portion of developers , games enthusiasts and other people who spend serious money on software etc..

Reply on news.ycombinator.com