Microsoft + Collabora Working To Map OpenGL/OpenCL Over DirectX 12

[Grim85]

Or dont they "we love linux" ?

Only when it sells CPU cycles in Azure

[mlau]

Apparently MS have money they don't know what to spend on. Let them start supporting Vulkan instead of DX12. That would be a proper investment.

At first glance this looks like a desperate attempt to keep DX12 relevant.

[CochainComplex]

At the time DX was faster than OpenGL and benefited of MS' backing but times may have changed. Online gaming services, if successful, may tip things in favor of cross-platform API's (VK) since many servers run on Linux (Google Stadia).

That is correct. But Microsoft is still pushing their DX - besides Xbox uses it too. That means Game Devs for Console have a rather easy way to port it to PC. So the pushing is still there and should not be underestimated. Dispite of Vulkan beeing on par with DX12 concerning performance.

[CochainComplex]

My two cents worth:

"In addition, we are taking advantage of Microsoft's experience in creating their D3D12 Translation Layer, as well as our own experience from developing Zink."

Yep, Zink, the existing Collabora (yep, same company, heck same dev) project that does OpenGL to Vulcan translation. Hmm, maybe there could be some cross value there... Later from the blog post:

"In addition, we believe this is good for the wider open source community. A lot of the problems we are solving here are shared with other drivers and translation layers, and we hope that the code will be useful beyond the use cases listed above."

Hmm, you mean things may be learned that have value elsewhere? You don't say.

I'll leave it at that, others have made additional points that don't need to be repeated. But yeah, "Microsoft bad" so there is that. And all those other dead-end projects (looking at you Canonical!) that are fun to b**** about and complain all but a waste. I wonder if any of those had any learning experiences involved that might have some value going forward. Naw, that could never happen!!!

Thats the point. The working power is deducted from Zink to this dead end ..lets call it proprietary opensource .....propensource Project.

[CochainComplex]

I never understand why people get upset by news like this. Opensource does not work like closed source. People seem to not understand it, even after decades and decades of opensource development. In opensource, people work on what they want. You like their work? Fine, you can benefit from it, no? Fine, don't use it. It is not like Microsoft funded Linus Torvalds to scrap the Linux kernel or anything, so why people complain?

I repeat because people are poor readers: There are devices that either don't have OpenGL/CL/Vulkan drivers at all, or have really poor native support, but they have great D3D12 support (because they are from Microsoft, duh!). There are 2 options, you either create a situation where a large portion of users do not have access to Khronos APIs (and thus make them less attractive to use for application developers), or you create a layer to run them over D3D12 (and thus make Khronos APIs more universal and thus more attractive to use). It is that simple. Why people complain about this? It is literally bringing support for Khronos Open Standards to devices that don't have support for those standards, or have poor support. It is something. And you are not paying for this, and you are not working on this. It does not affect you or Linux in any way, shape, or form. You just have to complain because "Microsoft is evil" (even though Opensource companies like Google are far more evil....)

Also, the prospect of contributing work towards making MESA OpenCL 1.2 compatible is really great and important and a great boon for Linux. Clover is the only way we are going to have universal OpenCL support out of the box, without having to resort to external libs, opensource or binary, with all their incompatibilities and headaches. This will drive more OpenCL 1.2 app adoption (since Linux userspace devs will have knowledge that the vast majority of desktop machines will have a baseline of 1.2 CL support. Perhaps we will see more desktop apps try to utilize it in various areas to improve the desktop, like how they utilize OpenGL these days to make composite desktops a universal thing.

Well I guess I'm also targeted by this comment.
So you are right - I have been thinking like this in the past as well. And yes there is some certain truth in it. But the issue is more abstract and not obvious at the first glance.

Let me give an example where open source gets "abused" even if it is legally flawless for corporate interests.

Windows Subsystem for Linux. At the first look it is cool. Yes Linux Applications on top of Windows - people will use Gnome, KDE on Windows and then it will get a broader community and will get more audience and devs. All that cool Linux tools usable on windows - great.

No.

Why is that? What is the intention. Microsoft especially their Server Environment suffer from the pressure of Linux and derivates in this fields.
How can one claim property back? Support Linux stuff on Windows. You will have both sides on one system. Because it is opensource Microsoft doesnt even have to pay one penny for all the cool linux tools - yes that is opensource, therefore Ok.

Let us reverse the case. Linux (Foundation/Community) tries to achieve the same on Linux. Native tools and the good tools of Windows.
Not possible/or difficult (see wine) because Microsoft is in control of the Windows environment and will not open it up.

So Microsoft will have _always_ the easier position in this fight. Microsoft has it easier to provide a flawless user experience in this scenario.
What does this mean -> people will stay or switch to windows if they have to use both sides of the world.

This can be easily determined by this comparison

Windows: Win Apps support: native + Linux Apps support: great
Linux: Win Apps support: ok + Linux Apps support: native

How would you decide just by logic? and who is going to profit at the end?

Keep in mind:
Corporate focus: profit / Opensource focus: community - Microsoft or any other company do only participate in "opensource" Projects if they see gain in profit or a better market position.

[log0]

Thats the point. The working power is deducted from Zink to this dead end ..lets call it proprietary opensource .....propensource Project.

You are free to pay Collabora to deduct more working power to Zink if you wish so.

[CochainComplex]

You are free to pay Collabora to deduct more working power to Zink if you wish so.

btw. ...this is on my "if I would be a Billionaire" List but there are some 0 missing on my bank account. Feel free to help me becoming one.
I promise I will put the money into opensource Projects instead of backing soccer teams.

[L_A_G]

Now this really had me stumped and wondering why they'd waste time on something like this...

Then I remember what Atari did back in the late 70s before they got bought out by Warner Communications. You see they had most of the companies who could, within a reasonable time frame, design a competitor to the Atari 2600/VCS working on projects that were never supposed to go anywhere. The actual purpose of these projects was to tie the companies up in exclusivity contracts so they couldn't design competing consoles for competing companies.

Warner of course saw this as a waste of money and killed all the dead end projects after they bought Atari, releasing them from the contracts and what followed was an absolute glut of new consoles that didn't end until the U.S console game market crash of 1983.

[CochainComplex]

Now this really had me stumped and wondering why they'd waste time on something like this...

Then I remember what Atari did back in the late 70s before they got bought out by Warner Communications. You see they had most of the companies who could, within a reasonable time frame, design a competitor to the Atari 2600/VCS working on projects that were never supposed to go anywhere. The actual purpose of these projects was to tie the companies up in exclusivity contracts so they couldn't design competing consoles for competing companies.

Warner of course saw this as a waste of money and killed all the dead end projects after they bought Atari, releasing them from the contracts and what followed was an absolute glut of new consoles that didn't end until the U.S console game market crash of 1983.

exactly. Besides it looks like a generous act to the most of the people (the opensource version of it). That is the sneaky thing about it.

[chocolate]

I never understand why people get upset by news like this. Opensource does not work like closed source. People seem to not understand it, even after decades and decades of opensource development. In opensource, people work on what they want. You like their work? Fine, you can benefit from it, no? Fine, don't use it. It is not like Microsoft funded Linus Torvalds to scrap the Linux kernel or anything, so why people complain?

I repeat because people are poor readers: There are devices that either don't have OpenGL/CL/Vulkan drivers at all, or have really poor native support, but they have great D3D12 support (because they are from Microsoft, duh!). There are 2 options, you either create a situation where a large portion of users do not have access to Khronos APIs (and thus make them less attractive to use for application developers), or you create a layer to run them over D3D12 (and thus make Khronos APIs more universal and thus more attractive to use). It is that simple. Why people complain about this? It is literally bringing support for Khronos Open Standards to devices that don't have support for those standards, or have poor support. It is something. And you are not paying for this, and you are not working on this. It does not affect you or Linux in any way, shape, or form. You just have to complain because "Microsoft is evil" (even though Opensource companies like Google are far more evil....)

Also, the prospect of contributing work towards making MESA OpenCL 1.2 compatible is really great and important and a great boon for Linux. Clover is the only way we are going to have universal OpenCL support out of the box, without having to resort to external libs, opensource or binary, with all their incompatibilities and headaches. This will drive more OpenCL 1.2 app adoption (since Linux userspace devs will have knowledge that the vast majority of desktop machines will have a baseline of 1.2 CL support. Perhaps we will see more desktop apps try to utilize it in various areas to improve the desktop, like how they utilize OpenGL these days to make composite desktops a universal thing.

Your point is worth considering.
Although, I believe it's one thing when FOSS crosses over to Windows to spread awareness and provide a viable alternative to proprietary tools and programs, and another when OSS is used to a) market MS' position as the good guys that they obviously aren't, while at the same time b) actively promoting and spreading non-standard protocols/APIs. There is one high-performance graphics&compute standard, and that's Vulkan.
Shiny new Windows 10X devices don't have support for Vulkan? It must be the customers' loss and ultimately MS' loss as they will have to either encourage OEMs to support proper standards (after all, they have shown themselves capable of such things, with the likes of Secure Boot), or they will lose said customers.

I like to think of this specific example as OSS prostitution.