What Should Valve Do For Linux & Open-Source?

[Kano]

@Figueiredo

Coreboot is hard to test on real hardware these days. The sample implementations for newer boards are only for evaluation boards which you can not buy. Ok, you could search some 2nd hand ones to play with it or just use qemu/kvm. I would really like to test it as you could even use linux as payload (or maybe grub). In theory uefi would be enough as well, in theory because current boards are too slow with POST. One interesting spec for W8 systems is that POST may not be longer than 2s. I really want to have got such a board because current retail boards are more around 10s for POST.

[gamerk2]

The ones that consider that a windows binary wrapped around in WINE is a good idea, you are delusional .

I use WINE for many years (and contributed to bug reports) and no matter the great progresses, it will be ALWAYS a crouch.

There clearly performance issues that will NEVER be solved, compatibility issues, installation issues, audio issues, and graphical issues....

I NEVER saw a game (except maybe really old) that perform same EXACT way than in Windows.

WINE is a good option....as in :last resort to play a game in LINUX.

Any dev to go with it to make a Linux "compatible" game is doomed to fail as for sales goes.

More or less my thoughts on WINE too. You are basically emulating parts of Windows, so there is a performance penalty that you will have to deal with. Its also impossible to create a DX wrapper, due to its tight integration with Windows. Only real way to make Linux ports is to convince devs to switch back to OpenGL, which they will not do until there is evidence that Linux users are willing to pay for closed source software.

Prove they can make money, and developers will flock to linux. Problem is, theres currently no proof that spending the time to make linux ports is financially viable.

[Hamish Wilson]

Prove they can make money, and developers will flock to linux. Problem is, theres currently no proof that spending the time to make linux ports is financially viable.

Humble Bundle and all those other Indie devs are not enough?

[johnc]

Humble Bundle and all those other Indie devs are not enough?

Probably not. It's nice for some "extra cash" but it's hard to say if the effort paid off. Not to mention a decent chunk of that Humble Bundle stuff was from rich people dropping huge donations.

Rage kinda speaks to the issue a bit... OpenGL, but no Linux port. If there was money to be made, they would've done it.

[Novan Leon]

"Whether you're a Linux gamer or not, it's to everyone's benefit that Valve's striking Linux work is steaming with greatness. Without the very best developers the Valve Linux cabal could be left for dead or ricocheting through a portal that has a half-life that is too short to make everything a reality."

Wow, the pun laden goodness of that last paragraph was staggering.

On-topic, I think Valve would do well to focus on a their own distro, something with their name on it, and focus all their efforts on the "Valve-Linux OS", whereby all other Linux distributions will benefit. This would give them a little more control over the OS environment and help them define standards for others to follow. Most importantly, it would create all manner of media buzz and help get people excited about Linux in a new way, something that would benefit all Linux users and give them more sway with companies like Intel or AMD. Sometimes the media buzz alone has enough value to effect change.

[johnc]

"Whether you're a Linux gamer or not, it's to everyone's benefit that Valve's striking Linux work is steaming with greatness. Without the very best developers the Valve Linux cabal could be left for dead or ricocheting through a portal that has a half-life that is too short to make everything a reality."

Wow, the pun laden goodness of that last paragraph was staggering.

All it needed was one self-link per word and it would have been a Michael special.

[nightmarex]

More or less my thoughts on WINE too. You are basically emulating parts of Windows, so there is a performance penalty that you will have to deal with. Its also impossible to create a DX wrapper, due to its tight integration with Windows. Only real way to make Linux ports is to convince devs to switch back to OpenGL, which they will not do until there is evidence that Linux users are willing to pay for closed source software.

Prove they can make money, and developers will flock to linux. Problem is, theres currently no proof that spending the time to make linux ports is financially viable.

You mean like prove XBOX was going to work? Do you know how much money was lost when MICROSHAFT jumped into the console wars? Now it's the defacto standard (360) and MS had to reach deep to make it happen.

Dev's don't give a crap what they write for, in fact, a few I know hate D3D's lock in so it's not all roses. Same with Apple when Blizzard released Diablo 2 for it... how much of the market did they have circa 2000? Quote from wikipedia "Statistics from late 2003 indicate that Apple had 2.06"

Diablo was a AAA title, no doubt and regardless of what the statistics say I would imagine that there is at least as many if not WAY more Linux users now as Apple had in 2k3.

So what was your point again? It just takes faith and you have MICROSHAFT and CRAPPLE fudding their minds. They know it can be profitable their just being pansy's about it IMHO... Just say well we only support official NVIDIA/AMD/INTEL drivers then they will be right where they are on Windows. Assuming they use OPENGL from the get go, which doesn't happen much.

[gamerk2]

Humble Bundle and all those other Indie devs are not enough?

Prove people are willing to purchase a $60 game, that requires a total overhaul of its graphics engine [DX to OGL port], and likely needs significant Windows API's to be replaced with Linux ones. This means more staff [Linux devs] and a significant financial investment for a very small part of the market, and a part of the market which isn't exactly known to be willing to throw money at non-open source products.

[gamerk2]

You mean like prove XBOX was going to work? Do you know how much money was lost when MICROSHAFT jumped into the console wars? Now it's the defacto standard (360) and MS had to reach deep to make it happen.

Heres the difference: People buy consoles to play games, which requires people to spend money. Linux is, frankly, not an OS who's demographic is known to splurge on games. Point being, if you develop SW for a console, there is the expectation it will sell if it is popular. On linux, that outcome is far less clear.

If Linux wants to be a major player in gaming, there needs to be a concerted effort to make middleware engines (*coughUnrealcough*) support the OS. Do that, and Linux becomes just as attractive as any console OS, and you remove a lot of the work needed to make a linux port.

[uid313]

Prove people are willing to purchase a $60 game, that requires a total overhaul of its graphics engine [DX to OGL port], and likely needs significant Windows API's to be replaced with Linux ones. This means more staff [Linux devs] and a significant financial investment for a very small part of the market, and a part of the market which isn't exactly known to be willing to throw money at non-open source products.

The investment is not only for Linux, because Mac OS X uses OpenGL too. As does Android, Wii, PlayStation, etc.

Also it doesn't necessarily mean more staff, just be smart and use cross-platform APIs from the start.

Heres the difference: People buy consoles to play games, which requires people to spend money. Linux is, frankly, not an OS who's demographic is known to splurge on games. Point being, if you develop SW for a console, there is the expectation it will sell if it is popular. On linux, that outcome is far less clear.

On the contrary, Humble Indie Bundle have shown that Linux users pay on average more than Mac and Windows users do for the bundle.
Linux users payed twice as much as Windows users!

If Linux wants to be a major player in gaming, there needs to be a concerted effort to make middleware engines (*coughUnrealcough*) support the OS. Do that, and Linux becomes just as attractive as any console OS, and you remove a lot of the work needed to make a linux port.

Unigine, Unity, id Tech and soon Source Engine are all available on Linux.