I'm just a kid that tries to cover PC building and gaming the best he can despite running on a beyond-destitute gaming laptop from years past (2008- five years!) and also talks about assorted Japanese fuckery like No More Heroes and Kingdom Hearts.
I'm also a freelance writer, have made quite a bit of money off of it ($1500 a year for a then-16 year old? Fuckin' sweet!) and I make a lot of jokes (too much) about penises and whatnot.
Keep your expectations low and you'll have a great time. Keep your expectations around the middle and you'll have a good time, but don't have high expectations around me because I am literally incapable of not disappointing you.
Let's cut straight to the point. Valve's been talking more about the Steam Machines than ever before. These steps started about a year back with the announcement of Steam's Big Picture Mode, and now Valve is on the verge of completing its advance to the living room with their announcement of SteamOS, the Steam Machines, and the Steam Controller, a strange gamepad/keyboard-mouse substitute hybrid that nobody except those that've actually used it really knows how it works.
But let's address these plot points in order.
Valve's made big claims- whether they realize it or not, that bought our expectations to match.
SteamOS is a Linux-based operating system that will rotate around Steam in some manner. That's all we really know about it at this point- Valve has yet to release this "specialized gaming distribution", and for all we know it could turn out to be a disaster. Like many other PC Gamers, I game primarily on Windows- but I've used Linux before, a Debian-based Distro called Linux Mint. I enjoyed the operating system wholly- after the initial setup and toying with the Bash Command Terminal, I was using an operating system that I truly enjoyed for the first time- but the lack of support for my favorite programs, such as Skype, and the sad state of WINE, PlayOnLinux and my own computer at the time forced me to stay on Windows 7 more than I would've liked to.
However, the loudest critics of Linux tend to be, as you may have guessed, people who don't actually use it. In fact, because of my disappointment with Mint, I was initially extremely skeptical about Valve's ability to do much of anything with the platform.
And that's where we're wrong. OpenGL is less resource-intensive and just as graphically capable as DirectX, and now console gaming is adopting the x86 architecture officially, with six of these consoles eight cores being devoted wholly to gaming. The consoles will be underpowered in no time- they already kind of are, but hardware optimizations will do incredible things with them, and for the first time, PC Gamers finally benefit from these as well, since these console games are finally being built for PC and then ported and optimized to consoles, rather than the other way around.
Even without Valve, WINE and PlayOnLinux have made significant advances in terms of their capabilities- yeah, you can't max out Crysis yet, but in some cases, you can run it. These advancements are without Valve's touting of a new Linux OS for people to crowd behind, and with the support of Nvidia and Valve's publishing partners, you're going to be seeing more purely for the Linux kernel and OpenGL itself. The momentous progress that has been made on Linux has been credited to two parties: the graphics vendors, AMD and Nvidia, and, more importantly, the developers that line the open source community. In the latter's case, these people aren't being paid for their work. Optimization and development on Linux will become much stronger and much more widespread if Valve makes a strong showing with SteamOS- and the people that'll advance that are both the open-source community and the programmers at various dev companies.
Linux is a viable gaming platform if publishers and developers buy into it. Valve's on some dangerous proving grounds here, and they'll need a lot more than their decade-old Source engine to pull it off.
Fortunately, that's what their new engine- which I think will be called Src3, since Source is actually the second Valve engine, the first being GoldSrc- will be for, and if they're capable of debuting this console-PC hybrid with, say some threequels and a new engine, they'll be making a strong showing for Linux, their Steam Machines, and, ultimately, the expansion of Steam.
However, PC gaming has long fallen victim to the difficulty of supporting hundreds of thousands of different configurations of graphics cards, processors, and memory amounts. Optimization on the level of a console for such a fragmented hardware market is basically impossible, but graphics drivers for modern video card series helps shrink the gap.
Even if Valve is capable of somehow changing the face of PC gaming (again), how on Earth could they repeat that task with Steam Machines, of various configurations, from various manufacturers?
Why, they're all from the same manufacturer. All the graphics cards are currently being supported by Nvidia, and all the CPUs are current-gen Haswell processors from Intel.
Valve has said in the past that these machines are going to come in three tiers- I'm paraphrasing, but they're basically Streaming, Gaming, and Dominating.Pardon the nerd in that link and let's move on: the Streamboxes are basically going to be cheap little things that existing PC gamers could use to stream their Windows titles from their bedroom to their living room. This brings up the obvious question of latency and compression, though on a decent home network the former shouldn't be a problem. The latter will probably see your games being displayed on your TV at something lower than the native resolution they're running, unless the Steam Machines debut some kind of super-efficient streaming mechanism or image upscaling. The other two tiers are basically just gaming PCs, and if SteamOS proves unsatisfactory, you could always install Windows or change your hardware.
DToid recently made some builds based on these specs, in case you want to buy those builds or estimate the final prices based on them. However, with all due respect, that article is extremely error prone. Not a single proposed build in it works- here's mine. I have nothing but love for DToid, and that's the reason why I made working builds that abide by Valve's specs- because I don't want people buying incompatible hardware or thinking the real Steam Machines are going to be hella expensive.
That aside, let's continue to the manufacturers.
If the manufacturers have to abide by Valve's baseline, it's likely that the Steam Machines will all be required to be touting 16GB of DDR3-1600 RAM, a GTX 600/700 series graphics card, and a Haswell processor. Graphics drivers for the GTX 700 series benefit the entire line of cards, in addition to the 600-series predecessors. Touting an i3, an i5 or i7 with 16 Gigs of usable RAM in all configurations means there is no hardware fragmentation.
Software fragmentation, then! Android faces severe hardware and software fragmentation! You can't beat one and still beat the other!
Android is fragmented because it's an open source project that the hundreds of different phone manufacturers can tweak as they please, while the phone carriers can delay the latest Android updates for each of these phones indefinitely. It's an open platform- and, personally, I love it- but hardware and software fragmentation plagues the consumer and the dev community alike in that situation.
In Steam's case, however, the client itself is constantly updated. In SteamOS' case, it'll likely be updated in the same manner that Windows Updates, Linux system updates or even console updates already work- by connecting to the internet and downloading the latest fixes and additions to the operating system.
This means that, a year into buying a Steam Machine, updates for your operating system won't mysteriously disappear. SteamOS doesn't face the same troubles that Android does- in fact, SteamOS and the Steam Machine demonstrate something thought impossible for PC gaming- a potential end to hardware and software fragmentation.
Sure, you'll have different tiers of effectiveness. Sure, you'll probably have to update once every four years. I don't doubt that. The weakest possible build of Valve's prototypes is a rough equivalent to the PS4, but even that will be quickly overtaken by the PS4's optimization capabilities.
No. The developers, especially if they get behind OpenGL and the Linux kernel, don't have to worry about that at all. OpenGL's cross-platform, too- even Windows gamers could see big benefits from this.
Finally, let's talk about our googley-eyed friend: the Steam Controller.
The dual-trackpadded elephant in the room.
What the fuck is that thing? What? WHAT?
It's primarily a keyboard/mouse substitute- but, according to Team Meat's writeup on the controller, which I linked to earlier in the article, there's also more to it than meets the eye. There's four triggers and two buttons on the back of the controller, the equivalent of dual analogs on the front, a touchscreen offering a theoretically unlimited amount of buttons in the middle, but at least two (three, if you count pressing the screen itself) at a time, and then four buttons around that touchscreen. But wait- according to Team Meat's write-up, the trackpads can be configured to use as...face buttons and a D-pad?
Let's assume the dual trackpads are somehow perfectly suitable for analog substitutes- maybe even better. I'm gonna bet on Valve here and hope I don't look like an idiot next year. If we assume that, then this controller has enough potential button configurations to serve the purposes of just about any PC game, RTS, FPS or otherwise.
But gamepad enthusiasts- myself included- know for a fact that games like, say, Arkham City play like shit on a keyboard and mouse, and they probably wouldn't be the most comfortable with that kind of control scheme on this controller, either.
If this controller can act as a 360 controller substitute, it may offer a solution to that problem as well. You know those buttons on the back, on each of the handles?
The 360 doesn't have an equivalent to those.
You're never able to use an analog stick and a face button/D-pad configuration on the same side simultaneously. You don't have two thumbs on each hand to do that with, though gamers that've played on the PSP and have had to use the D-pad as camera control can definitely vouch for how uncomfortable that would be.
What if, normally, the trackpads could serve as analog sticks, and if you decide to hold down that button in the back, they could switch to serving as normal buttons? Or, if the touch input is seperate from the five button inputs available on each trackpad...you could be using analog movement one moment and then pressing a particular point of your trackpad like it's a D-pad the next.
There's a lot of ifs, ands or buts in this article, but I don't doubt Valve and I don't think you should, either. Historically, that doesn't turn out very well. I'm sure there's plenty of people who wish they could've tapped into the PC gaming platform the way Valve did. (No, we're not talking about Origin.)
The point is, I believe SteamOS will prove itself a worthy gaming platform, that the Steam Machines various configurations will serve as both a unified architecture for developers to work for and a cost-effective route for PC gaming no matter the budget of the buyer, and that the Steam Controller will actually not be a complete and utter abomination.