Valve hired their first wave of Linux developers, hiring more | Hacker News Reader

Valve hired their first wave of Linux developers, hiring more

phoronix.com

216 pointsshazow87 comments

This will be welcome by me, steam on Linux that is.

I've bought a number of games from steam when they have been on sale, games that have a Linux port available.

Of course I am unable to run them without buying the Linux copy separately at present but once this is implemented I will already be able to play a fair cross-section of my games on Windows and Linux.

Of course this could be a net negative for my productivity.

Linux developers doesn't necessarily imply there'll be Steam for Linux Desktop™. It may be that Valve is looking to finally build their Steam Console and base it on Linux.

That's true, but this article seems to indicate that it will be there for the regular "desktop".

http://www.phoronix.com/scan.php?page=article&item=valve...

It is of course possible that they are planning a console and that this is why they are focusing their efforts.

If they do develop a Linux based console then hopefully it will still make sense for them to support "proper" Linux as part of that.

I think that if they support 'proper' Linux might also depend on the architecture of the console. If it's PowerPC or ARM, then we might not see x86 desktop support.

I'm going to take a guess that Valve is hedging their bets on a Windows 8 flop. I think it is very possible that Linux and OS X will become the two dominant desktop platforms over the next 5 years.

Valve can not really enter the mobile marketplace in a meaningful way. Both Microsoft and Apple will likely try to put the squeeze on Valve with their own respective OS marketplaces. (I think neither will successfully be able reproduce Steam's high quality user experience and extras.)

Perhaps they want to port to platforms like the new chromebox (or a future variant).

I think we are about to see a divergence in desktop/HTPC/semi-mobile platforms.

For example, ChromeBook/Box , Raspberry Pi , Android , Ubuntu Devices and I'm sure many others as time goes on.

it seems reasonable that people will want to run games on these platforms beyond what is currently possible just with JS/webGL.

Now think about it, what do all of these plaftorms have in common?

Another point to bear in mind , is that one of the significant advantages Steam has over OS specific marketplaces is that it allows you to sync the same games over different devices.

For example it would be nice to be able to buy a game for your Windows PC and then be able to also run it on your HTPC or Tablet device which may well be running non Windows OS.

It's a bit more complicated than that. Valve is hedging their bets against a Windows 8 win.

Microsoft is playing a longer game than most people can see. So is Valve.

I think that if Linux becomes the dominant desktop platform, it will be a new form (i.e. not based on Gnome or KDE). Just a (biased) hunch, though.

Ubuntu is positioning itself for that possibility. They've been playing the long game, and it's still not finished, but they've broken from Gnome, broken from X.org, and with the Software Center, are really steering away from the terminal. Of course, they've pissed off a lot of Linux traditionalists, but that could be a net gain for the desktop market.

No, I this is Steam for Linux. The Steam Console was just a rumor, and that's all.

I've had an employee (friend) verify these facts to me.

steam for linux is coming http://www.phoronix.com/scan.php?page=article&item=valve...

There's a lot of overlap between Steam/Linux and Steam/LinuxConsole so it wouldn't surprise me if both strategies were being run in parallel.

For some games, you can just run Steam in Wine, and the game in Wine, and everything runs out of the box. Check the Wine documentation for each game before buying to ensure it works.

Yup, Portal ran perfectly for me on Ubuntu ( 9.10 or 10.04 iirc) that way a few years ago. Was brilliant.

Even though I don't game anymore, this still makes me very happy. A big step for Linux on the way to becoming a first class citizen in gaming, one of the last big moats to its general acceptance.

Sort of off-topic, is there a reason Steam on windows is so unresponsive? Mac, too, but I think that's because it's a port.

It seems to me that Steam is not written in native windows code and that's why it runs so sluggishly. Some say they have no idea what I'm talking about, while others agree. What does HN think?

They use their own UI toolkit called VGUI, the same one they use in HL2 and their other games (https://developer.valvesoftware.com/wiki/VGUI_Documentation). So, there is a layer on top, but in terms of rendering it should be all native. In their games, it uses Direct3D calls, and in Windows I would guess it uses GDI+ or something in the DirectX family.

Edit: Just to be clear, I would consider the VGUI layer to be pretty lightweight. There is no virtual machine for any language runtimes. It is purely a layout engine, with some very basic styling defined mainly by the rendering of images and gradients.

Outside of rendering, from my own experience with using it, there's nothing in particular about it that makes it slow, except that it has no built-in asynchronous facilities in place for you to use. You have to make sure you never block the UI thread.

This is really just a guess of course, but I would pin the slowness on too much IO waiting (ie. poor thread design). To be fair though, Steam in most cases has stayed pretty responsive.

I would rather put my money on not threaded well. The application thread seems to get blocked whenever it's chugging on something important, which locks up the UI.

Define "so unresponsive".

Recently (less than a year ago), I had a switch in hardware from a Athlon X2 3200+, 4GB RAM machine to an i5-2500K, 8GB RAM and 2 SSDs. Sure there was a noticeable difference in responsiveness, but never on either machine would I consider Steam "so unresponsive".

My last machine was an i7 920 with 12GB and Intel 80GB SSD; I now have an i7-3770K with 16GB and Samsung 256GB SSD; and Steam has always been fairly unresponsive.

Specifically, it's most unresponsive on the "preparing disk space" page when installing games. It takes multiple seconds for no apparent reason. There's no high CPU usage and no high I/O usage. It's also very slow to start, and very slow to let me play a game.

My theory is it's due to some degree of online activity, that it's gated on my internet bandwidth. Since I only play single-player games, that's a huge annoyance to me, and I often don't start playing games because I know I'll have to sit through Steam startup for several seconds.

Steam does have a very slow boot time, but the expected behavior is that Steam is always running in the background so I can't imagine it's a priority. I can't fathom a ~3 second wait when installing and launching a game being a real problem. Minor annoyance sure but that's less time than you'll be forced to see stupid splash screens for!

I just quit Steam (so it was fresh in cache, never mind relying on SSD etc.), and started a Steam game. It took 20 seconds before the game launched (not playable, just started).

That's massive. Pretty much any of my non-steam games will be launched and playable within 10 seconds.

(I also hate unskippable intro videos; I'm fairly sure I've collectively spent much longer searching for ways to delete / skip such videos than I have ever lost watching them.)

With Steam already running the moment I double-click an installed game from my list it's taking about 2-3 seconds before they pop open full screen. If you're taking 20 seconds then it may or may not be Steam's fault, but it certainly isn't an intrinsic property of the software.

Sure, but my point is that I tried it without Steam running (most of the time seems spent on logging in to my account). I don't run Steam by default - it'll be updating, downloading patches, popping up notifications and all sorts interfering with my daily life.

That is neither the expected nor the standard use case. If you use a product in a way it's not designed to be used you shouldn't be surprised when it's behaves in a non-ideal fashion. Notifications and auto-update can be disabled.

That's completely unacceptable - the idea that I need to keep a background app running at all times just so I can play a game once or twice a month. Imagine if all apps took such an arrogant approach; our machines would barely run under the weight of the cruft. I similarly avoid Intel installers when using their drivers because they also like to install and keep running a half dozen services and widgets. Same thing with iTunes - I use a shell script to start it up so it can start up all the extra services and kill them later after iTunes exits.

There's a lot of entitled software out there that arrogantly assumes sovereignty over your machine. I don't accept it.

Steam performance seems to very inconsistent.

It often runs like a dog on my Core2Quad but is usually butter smooth on my P4 laptop.

Right Click on a game -> Properties -> Set Boot Options -> "-novid"

For all my Source games: "-novid -w 1920 -h 1080 -window -noborder"

As far as I know, it uses Webkit for the entire UI--and possibly a rather old version of it.

Yes, that is true to an extent. For the web portions of the UI like the Steam store, it uses the Chromium Embedded Framework: http://code.google.com/p/chromiumembedded/

Steam itself works great; the only time I see problems with it is when the flashplayer plugin is involved.

They desperately need to drop their reliance on flash.

Once they do that, I think it would vastly improve.

Agreed, I nuked the flash player on all my osx/linux/windows machines (chrome deals with that decently enough) and now I can't view videos in steam.

Parts of it are served from the internets.

Steam works beautifully under Wine on my Ubuntu laptop (A Thinkpad T420).

I've never ran Steam under Wine but I have played quite a few games with varying degrees of success. My question is since Steam runs beautifully for you under Wine, does that translate into the games themselves running beautifully as well or is it still hit and miss?

Purely depends on the game. A good guess is that as long as the game runs on DirectX 9, it should run reasonably well. That should be most games, except for some of the more recent ones.

Great news! Honestly, the only software that will be holding people back from switching fully to linux now will be creative tools (audio, 3d + 2d design, video editing). I think this can be solved by us. If we show them that there is a demand for these tools on linux someone will move to supply. Open operating systems provide too many advantages to ignore.

Well when it comes to 3D Linux is actually well off due to it's strong position in SFX, you have the latest Maya, XSI, Mudbox, Renderman, Houdini, Lightwave, 3D-Coat etc available in native Linux form.

And as others have stated there's also great free solutions like Blender.

Thanks for informing me of this, I honestly didn't know they worked in native linux.

This is what XSI has to say about Linux - "Softimage 2013 is also capable of running on other configurations, such as boutique distributions of Linux. However, enumerating systems that are not tested and cannot be supported or that fall below the requirements for a productive user experience is beyond the scope of the online certification charts.".

Lightwave isnt available for Linux either.

Ah, I was under the impression that Lightwave 10 was released for Linux last summer, granted it's not really on my radar. Looking at their site now showcasing version 11 there's no mention of Linux so it seems you are right. Still, Wine's AppDB page for Lightwave mentions a Linux native Lightwave 10 CORE under development.

Yeah, the latest Lightwave runs on Linux. I installed one of the Beta's a few months ago and it ran OK, though it didn't include everything the windows version had. I still use Lightwave via Virtual Box.

Agreed, and with Adobe and Autodesk monopolizing creative software, I think we'll see more and more of this. Blender has been doing spectacularly with 3D to the point that Autodesk must be incredibly worried, and I expect we'll see more Photoshop contenders as projects like GIMP start to answer the pleas of its on-the-fence users. See the GIMP roadmap: http://wiki.gimp.org/index.php/Roadmap

Edit: And I expect as browser technologies continue to expand, we'll see more web-based competitors. In particular, with WebGL, I'm surprised we haven't seen things like all-GPU image editors to compete with the slow giants we have to deal with now.

Also in the video editing space Lightworks might prove to be an acceptable (if not even better) alternative to Avid and FCP.

In the mid-2000s, Macromedia started an experimental port of Dreamweaver (I think) to Linux using Wine. The project was abandoned because the market for commercial Linux applications was seen to be too limited for the difficulty of maintaining the port.

Krita (www.krita.org) is also very great piece of software recently, sometimes called "best-kept secrets among Linux applications". See this blog-post from professional illustrator http://www.davidrevoy.com/index.php?article98/krita-2-4-beta...

Some of the best tools in the industry (Audacity, Blender, Inkscape, others?) are already free, open source, and native on Linux.

None of these you mention qualify as being "the best in the industry". Neither do Gimp and Kdenlive or any other video editing software on Linux.

By whose qualifications?

I much prefer Inkscape over Illustrator for my vector art. I much prefer Audacity over... actually I'm not sure what's considered the direct commercial competitor here for post-processing audio. I mildly prefer Photoshop over Gimp for graphics design and tablet drawing, but after almost a decade of using each, they're practically interchangeable for me. I mildly prefer Blender over 3DS Max for making game models (though I admit it has been a while for me—I hear the recent Blender releases have amazing improvements). And that's not even taking into account the difference in license costs.

I'm not claiming to be an expert in any of these industries, but I don't think I'm completely naive here either. I can imagine there is a strong majority of experts who would never considere free/open source alternatives to things they've been using since they graduated their respective art school and paid thousands of dollars in licensing fees. I feel they're missing out, even though I'm biased towards open source.

P.S. I said "some of the best."

actually I'm not sure what's considered the direct commercial competitor here for post-processing audio.

Audio is an interesting market. Sound Forge comes to mind, but based on the time I've spent in studios, I'm going to speculate that professional audio is either done with hardware, or really old software.

The particular example I'm thinking of was a MacOS9 computer (in 2009 or so). The tech told me two things kept him back; he had all the plugins he needed for the software he used (and that many were not available in newer suites), and something about realtime recording.

Most commercial studios would be running either Pro Tools, Cubase/Nuendo or Logic. There is nothing in the open-source world that comes even remotely close to rivaling any of these packages at the moment.

Ardour is probably the most similar Open Source project.

@s_kilk: in random order:

Roadmaps: http://tracker.ardour.org/roadmap_page.php http://ardour.org/development/post3.0

MIDI: those Atari apps you're thinking of didn't do audio.

General utility: perhaps you need to talk to various people who do, in fact, use Ardour instead of ProTools. As for the comparison with Reaper - Reaper is a very impressive project, but if you try it for pro- usage rather than home-style recording, you will very rapidly discover quite a lot of basic issues that they have not addressed, but were being done correctly by Ardour 8 years ago. I look to Reaper for good ideas and impressive development speed, but they really are focused on the needs of a different type of market segment than I really pay the most attention to.

It is true that there is no single DAW that is to everyone's taste, and Ardour is probably a more "acquired" taste than many others. At this time, I would say that more or less no DAW can easily "replace" any of the others. You can't really replace ProTools with Nuendo either.

jsprinkles: we distribute binary copies of Ardour from ardour.org that run on any Linux distribution (just about - we discovered a wrinkle that interferes with this on a few platforms, and the next release(s) will correct this).

disclaimer: I am the lead author of Ardour.

Ardour is a cool project but again, it's not remotely close to Pro Tools, Logic, Cubase, or even REAPER. It's like comparing a military aircraft carrier with a canoe, sure they both float on water but they're not even in the same league.

I've kept up with Ardour progress over the past 6 years and it has consistently gone nowhere while the major players innovate continuously. The assertion that Ardour has a chance at competing with the professional packages is absurd given that they've been promising basic MIDI support for years now, a feature that was present in software packages on the Atari. Add to that the fact that the project doesn't even have a roadmap for future development and it all looks pretty grim.

My previous career was audio engineering, and I can't help but laugh when the FOSS crowd insist that Ardour and Audacity are capable of replacing the industry leading software packages.

Having used both, Ardour has a long way to go before it's remotely close to Pro Tools, even SE.

I think you might see interest in the platform develop if there weren't dozens of distributions to worry about. The distributions hamper Linux development, because you have dozens of different variants to try on, all with different versions of things and different bugs. As another commenter said, I bet you could get something Pro Tools quality if there was a consistent Linux-based desktop, not four or five frontrunners all doing GNOME differently.

I don't disagree. I still do all my music using Sony (formerly Sonic Foundry) ACID, despite being a nearly full-time Linux desktop user.

Sadly, some people refuse to work if they aren't using photoshop/illustrator/autocad/etc. But yeah, I'm with you, I use gimp and inkscape but I'm no professional.

It's not just the tools, but all the additional plugins that are built on top of those tools. Professionals don't just use Illustrator and Photoshop.

A perfect example are front-end developers. It can be tough for anyone to make a living in that line of work without being able to edit/open photoshop and illustrated without having to configure wine.

I truly believe Adobe holding back their products is one of the largest roadblocks linux has to get widespread adoption amongst the design/developer community.

I'd be surprised if Adobe isn't at least looking into this, especially considering their sometimes sour relationship without Apple.

A lot of the more "arty" people I know would consider giving them a Windows computer to use as a personal insult and would probably be more open to switching to Linux than Windows.

Getting arty people to use desktop Linux could be a huge net positive if they decide to help "pretty" the place up a bit.

It's certainly nice when UI/UX people run GNU/Linux distributions. Daniel Foré is in charge of Elementary OS [1]. Because he's a designer, the project already has a proper set of Human Interface Guidelines [2], a set of uniquely designed (and easy to use) core applications [3], and a beautifully coherent look to the entire OS and website.

[1]: http://elementaryos.org/ [2]: http://elementaryos.org/docs/human-interface-guidelines [3]: http://elementaryos.org/discover

I love Linux for general hacking, but not having Adobe tools is a huge blow when it comes to working on a design. It's almost insurmountable.

Exactly this, I'm a designer, and I like the look of ubuntu, but it's no macOS, it's hard to see an advantage that linux has over macs apart from the price - but designers need fast (expensive) machines anyway. No Adobe is the reason I can't even contemplate a switch.

I love blender though, my 3D tool of choice so at least 3D artists are well catered for.

Professionals have spent considerable time learning their tools (PS, Illustrator...), hence while these tools are not available on Linux they will always refrain from shifting to Linux and that is quite understandable. After all every need their work done best possible way with least obstructions.

> Some of the best tools in the industry (Audacity, Blender, Inkscape, others?) are already free, open source, and native on Linux.

Best to whom? None of the listed ones are used at all by professional game studios.

I think presence of mainstream CAD software is very important as well. I have seen many people not willing or not being able to shift to Linux only because of CAD.

I think that you need to specify which type of CAD software you have in mind. CAD software for Semiconductor designers have an incredibly strong presence on Linux.

I think in the domain of Civil Engineering, Architecture, Mechanical Engineering. My dad is a Civil Engineer and I know stuff he needs doesn't exist on Linux.

I meant that too. Basically when adobe and autodesk make their software linux compatible we will see a huge shift to linux. The amount of money saved by businesses on a whole would be huge if switching to linux was an option.

So that thing that this very same site has insisted is in full development and is near ready to go... is going through a hiring round for devs to build the thing? Right.

I'm so sick of these articles. Ever since the Mac version came out this site and other Linux focused blogs have insisted Steam for Linux is minutes away. Every single time they've been proven wrong or miss their timeframe.

It probably is being built, but the last person who has legitimate up to date info on the project is Phoronix. They seem to just be making things up.

IIRC they actually went to valve's office and showed screenshots of source engined games running under Ubuntu natively.

I wouldn't be surprised if they have it "90%" ready-to-go and just need some serious domains experts to do the last 90%. Not to mention for bugfixing , updates down the line as well as maybe porting some of their new tech (HL3 engine?) to Linux.

Phoronix seems to be playing a kind of Pascal's wager here. There is no way that it can be definitively proven that Steam will never arrive for the Linux desktop, but if it does then they can feel smug satisfaction that they knew all along and nobody believed them.

I don't use linux as my desktop OS for quite some time (now it's windows and osx, used to have Redhat and later fedora workstations). My question is, is it possible to target 'linux' in general as a single platform at all? I thought heterogeneous nature of linux distros is what would hinder that effort. Something like when you buy certain graphics programs, you get to have only this and that linux variant with this and that software on it in order to run it.

You don't need to ship distro-specific packages. You could ship a tarball with a statically linked executable. Take a look at the Humble Bundle, for example. Every game published in the bundles is cross-platform, and work great Linux regardless of distro choice.

Ah, never thought about that. As good old DOS games were, everything contained within program/directory itself. Interesting, makes using some of the external code a bit harder due to licenses though.

In the 'enterprise' market, targeting 'Linux' means targeting RHEL. I don't think that would be the way for Valve to go though. If they are really going to release anything to the public for Linux I imagine they will just officially target Ubuntu.

Ubuntu is focused on delivering a great desktop. I'm not a fan of Unity and find that to get the same performance I used to have, I have to run my games on Unity 2D. Regular Unity (3D) eats a lot of my games's FPS.

If Steam picks Ubuntu then I hope they work with Canonical to get Unity to perform as fast for games as Unity 2D (and Gnome 2).

That should probably be as easy as providing a way for games to tell Unity to turn off compositing, right?

I'd imagine that it would be some serious work to see whether just turning off compositing would be enough and what effects that has on running non-game apps that are in the middle of using compositing.

Windows does this already for games that enable it. Turns off desktop effects when it goes full screen, then reloads it when the game exits.

Will this finally mean linux becomes more and more mainstream for commercial games and applications? I actually hope so!

Thank you Gabe and Valve, for bringing things a step ever closer.

Does anyone else think this has little to do with 'bringing Steam to linux' and a lot to do with SteamBox, the rumored console Valve is working on.

I have a Dell Vostro with an ATI 6600m card running ubuntu 12.04.

Putting "echo OFF > /sys/kernel/debug/vgaswitcheroo/switch" effectively boosts my battery time from 1 hour to 3 hours.

What I'm seriously hoping from Valve is not a game per se, but a linux distro (or a standard) that is compatible with my drivers, multimedia devices (dual hdmi monitors - I'm looking at you) and audio/video.

I would pay serious money for that kind of a distro.

a linux distro (or a standard) that is compatible with my drivers, multimedia devices (dual hdmi monitors - I'm looking at you) and audio/video

yes, that would be great! Major distros (I use Kubuntu) are understaffed and concentrate on getting the new features from their upstreams integrated, not on making sure fancy hardware is supported. It's a question of resources. Although I think Valve is less likely than a hardware vendor to have a business reason for doing this work.

I don't think Yet Another Desktop Distro is really the answer to our prayers.

If fixing your laptop battery is as simple as that one liner , perhaps your can write a script that detects hardware at startup and runs that line if appropriate and then submit it back to Ubuntu?

The problem of Valve is their usage of DRM. I won't use them just because of that, whether they release Linux client or not. I.e. it's better to run a non DRM-ed Windows game (let's say from GOG) on Linux using Wine, rather than native one from Valve but with DRM.

But their entering to the Linux scene in general is good, since it might encourage others to do it as well, and eventually non DRMed distributors like GOG will push for more Linux games too.

Oh, thank Gabe! I've been waiting for this for a decade now!

At first I read valve FIRED their first wave of linux developers...

Reply on news.ycombinator.com