In a recent poll, less than one percent of Steam users are using Linux and Unbuntu combined.  If we throw in Mac we get 3.5% using that platform, most everything else is Windows.   Window 7 with DX 10, Windows 8 with DX11, Windows 8.1, etc., it is still all sorts of different kinds of Windows.  Windows might not be the best at everything, but it does quite a good job a gaming, working with all sorts of hardware, and different kinds of input, and developers like it.  And soon, it much of it will be updated to Windows 10 and many to DirectX 12.  This will add game DVR fuctions to everyone's PC (Like on the Xbox One) and other cool features.

Right now many people are requesting Half Life 3 and many other games that Valve is known for, but rather than work on those games instead they are developing their own operating system.  Valve is developing SteamOS, which is a gaming system - like Windows/DrirectX that is going to be used to challenge Microsoft as a platform.  The problem is they need outside vendors, i.e. 3rd party developers to make any real dent.  Now if it works say 10-20% better than Windows/DirectX, then I could see that actually happening.  But it seems very unlikely and that's a rather low bench mark.  Compounded by the fact than so few people run Steam on anything but Windows.

So they really have their work cut out for them.  It's a huge gamble and may pay off well, or be a colossal waste of money, time, talent, and man power. 

What, if anything, do you think Steam could do, or specs they need to hit, to make SteamOS a real contender for gaming, developing, and control over the living room?

Here is the survey for the curious: http://store.steampowered.com/hwsurvey

The question is moot. Valve does not care for wide adoption. It's just another option.
SteamOS is really only good for Steam machines that are dedicated gaming machines. Having SteamOS as main OS is just not viable for the vast majority of PC users.

Exclusives

Make it really easy to install and dual boot with (so people can have windows/linux + steamOS).

Gamle everything on Vulkan (the opengl version of mantle).

*invest in makeing sure vulkan is supported by the most popular game engine designers.

So that games can be ported to a vulkan version easily.

This, they're just trying to grow the steam ecosystem by offering different options such as steam machine, steamOS, steam controller, steam link...

The problem they face:

1) they cannot compete on price.
2) PS4 has a pretty good software ecosystem.


issues:

1)You cant build a pc currently thats faster than a PS4, but costs less than a PS4.

2)Since steampc's vary in components, you ll never reach the same level of optimisation as the PS4 (you need a stronger machine to do the same thing the PS4 is doing).

3) with a Consol, you put in the game, and play. You dont have to tinker with game settings to find what works.


The way I see it those are the main issues.... and their kinda big.

They need to appeal to the upper end of the gamer market instead.
Make crazy expensive SteamOS pc's (in terms of hardware) that can do what no other machines can do.

1) Steam games are much cheaper than PS4 games, so overall, I think it's much cheaper to play on SteamOS in the long run.

2) Steam (not SteamOS yet) is a much bigger ecosystem than PS4's. + no problem of backwards compatibility, you can change your hardware and still play your games.

1) You can't build a pc currently that's faster than PS4 for the same price YET. But you will, and you can already build a much more powerful PC than a PS4.

2) Valid point, but you're already able to build a much more powerful PC, so...

3) That's what Steam Machines / SteamOS are about.

Lol. Most gamers can't even build their own pc, they will certainly not tamper with dual boot. Also dual boot is very inconvenient if you want to switch between games and your windows applications. It's just not viable. They would have more success with a Steam branded Windows.

Vulcan is not different enough from DX12 and both are useless for most gamers anyway because only a handful use potato CPUs. The performance argument is irrelevant anyway because most people just don't care. Especially not as much as to change their OS.

The biggest problem, in my opinion, is a technical one: linux. Valve has to fix the mess that Linux is in terms of gaming. Different renderers, different window systems, different gl-versions, driver status with performance penalty... just to name a few.
If you want "serious" gaming besides 2d-games and basic 3d games like tuxracer, you have to mitigate a lot of existing problems.

But you will have the option of dual booting.  Which works, but some people find it annoying.

So just boot to SteamOS to play Half Life 3 whenever that comes out?  They really want other developers making games for it - do you think they can, or is it already too late?

So why even bother?  They could put that time, effort and talent towards making new games instead.  Should they just be an Gaming company and not try to be an OS/Hardware company?