During the Microsoft Build 2020 developer conference, Microsoft has raised a number of eyebrows at their Linux plans. We've had a lot, and I do mean a ridiculous amount of people emailing in and messaging across various places about Microsoft. So, to get it out of the way and provide you a place to comment, here we are.
Microsoft put up a developer blog post titled "DirectX ❤ Linux", which is a nice bit of PR bait. In reality, it means nothing for the standard desktop Linux. It's focused entirely on the Windows Subsystem for Linux which Microsoft tightly controls and DirectX itself remains firmly closed source. Not only that, this current implementation relies on pre-compiled user mode binaries that ship as part of Windows itself. Right now it seems to also be focused on CUDA and AI / Machine Learning, however, they also announced Linux GUI applications will eventually be supported on WSL as well.
A Microsoft developer even said on the Linux Kernel mailing list, that there's "no intent" to have people coding for DX12 on Linux. Although another developer also said they "consider the possibility of bringing DX to Linux with no Windows cord attached". That's just words for now though. I wouldn't read much into it.
That's not all, they also announced the Windows Package Manager under an MIT license, which works much like the ones on Linux do in terminal. Better late than never.
Going even further, Microsoft also announced .NET MAUI, an "evolution" of the Xamarin.Forms toolkit which Microsoft said "supports all modern workloads" which once again did not mention Linux anywhere. However, to be properly clear, at least .NET MAUI should work on Linux like Xamarin.Forms but be entirely community supported (as noted on GitHub). Oh and Maui is already used—oops? It's GVFS all over again.
First they embraced Linux doing away with the Ballmer era of "Linux is a cancer", now they're extending a branch saying they were "on the wrong side of history" with open source and now they continue the extending. How long before extinguish phase starts (EEE)? Don't be fooled about Microsoft's stance and their aim here, it's not because they love Linux. They're going where the developers are to continue pulling people to Microsoft services. Nothing more.
If any of it concerns you: I hope you put that energy and effort into continuing your support of the Linux desktop. Help it to grow and prosper. Support your favourite distribution, your favourite application and/or game developer by throwing some money at them.
Article taken from GamingOnLinux.com.
QuoteHow long before extinguish phase starts?
Never. Not because they don't want to, but because the importance of Linux as a whole is way bigger that than the worst-of-both-worlds WSL system can ever be. It simply doesn't matter in the grand scheme of things.
Quoting: EhvisNever say never when it's Microsoft. Changed in some ways, not at all in others ;)QuoteHow long before extinguish phase starts?
Never. Not because they don't want to, but because the importance of Linux as a whole is way bigger that than the worst-of-both-worlds WSL system can ever be. It simply doesn't matter in the grand scheme of things.
You could argue that trying to pull developers from Linux to stick with Windows, and just use WSL for ML/AI/CUDA is part of an extinguish strategy. It is when you think of it quite literally, as keeping people on WSL where they control it. All depends on your point of view and what we're referencing when we say extinguish.
Last edited by Liam Dawe on 20 May 2020 at 9:14 am UTC
Nevertheless May 20, 2020
Quoting: NeverthelessNext variation: "I love goose, says the fox." ;-)
and he says it without lying, he just means something very specific by that that the goose would not like! :)
Quoting: Liam DaweQuoting: EhvisNever say never when it's Microsoft. Changed in some ways, not at all in others ;)QuoteHow long before extinguish phase starts?
Never. Not because they don't want to, but because the importance of Linux as a whole is way bigger that than the worst-of-both-worlds WSL system can ever be. It simply doesn't matter in the grand scheme of things.
You could argue that trying to pull developers from Linux to stick with Windows, and just use WSL for ML/AI/CUDA is part of an extinguish strategy. It is when you think of it quite literally, as keeping people on WSL where they control it. All depends on your point of view and what we're referencing when we say extinguish.
But the landscape is different. EEE applied to things microsoft could control in some way. The battle fought against Linux was lost long ago and the foothold that Linux has in the total market is way too big for EEE to work. Sure they can do the first E. They can try to do the second E, but it is doomed to fail because it won't find significant adoption. Which means the third one is out of the question.
To me the whole WSL thing is a curiosity. They made a Linux kernel work on Windows in a VM like manner, but what is the point? To have your Linux service go down with a windows update? Making things dependent on two systems instead of one is a guaranteed way to make it less reliable. As a whole WSL actually looks the same as Stadia. Sure, there is a narrow niche that might see an advantage in it and would use it, but it is not objectively better and will therefore be ignored by the big crowds.
Quoting: EhvisAn interesting viewpoint. I'm curious then to find out how much adoption WSL has seen, if they made a bigger and better version of it with WSL 2 and now they're expanding it again - does that not mean the opposite of what you said? It must be getting a reasonably good adoption rate for MSFT to pour resources into it?Quoting: Liam DaweQuoting: EhvisNever say never when it's Microsoft. Changed in some ways, not at all in others ;)QuoteHow long before extinguish phase starts?
Never. Not because they don't want to, but because the importance of Linux as a whole is way bigger that than the worst-of-both-worlds WSL system can ever be. It simply doesn't matter in the grand scheme of things.
You could argue that trying to pull developers from Linux to stick with Windows, and just use WSL for ML/AI/CUDA is part of an extinguish strategy. It is when you think of it quite literally, as keeping people on WSL where they control it. All depends on your point of view and what we're referencing when we say extinguish.
But the landscape is different. EEE applied to things microsoft could control in some way. The battle fought against Linux was lost long ago and the foothold that Linux has in the total market is way too big for EEE to work. Sure they can do the first E. They can try to do the second E, but it is doomed to fail because it won't find significant adoption. Which means the third one is out of the question.
To me the whole WSL thing is a curiosity. They made a Linux kernel work on Windows in a VM like manner, but what is the point? To have your Linux service go down with a windows update? Making things dependent on two systems instead of one is a guaranteed way to make it less reliable. As a whole WSL actually looks the same as Stadia. Sure, there is a narrow niche that might see an advantage in it and would use it, but it is not objectively better and will therefore be ignored by the big crowds.
Quoting: EhvisBut the landscape is different. EEE applied to things microsoft could control in some way. The battle fought against Linux was lost long ago and the foothold that Linux has in the total market is way too big for EEE to work. Sure they can do the first E. They can try to do the second E, but it is doomed to fail because it won't find significant adoption. Which means the third one is out of the question.
How it is lost? Microsoft still has monopoly in the personal computer market and a very sizeable part of the server market. (Don't underestimate Windows Server, None of the top 100 website use it but it still widely used thanks to lobbying in the schools of system administration). Microsoft Office is one of the most used software in the world (not on Linux) and there is lobbying from Microsoft not to make available some pro software on Linux (like Catia, Solidworks or Abacus).
Quoting: Liam DaweJust a sidenote: If this is some kind of a long-term strategy of MS's, such a thing as an immediate adoption might not be really important.Quoting: EhvisAn interesting viewpoint. I'm curious then to find out how much adoption WSL has seen, if they made a bigger and better version of it with WSL 2 and now they're expanding it again - does that not mean the opposite of what you said? It must be getting a reasonably good adoption rate for MSFT to pour resources into it?Quoting: Liam DaweQuoting: EhvisNever say never when it's Microsoft. Changed in some ways, not at all in others ;)QuoteHow long before extinguish phase starts?
Never. Not because they don't want to, but because the importance of Linux as a whole is way bigger that than the worst-of-both-worlds WSL system can ever be. It simply doesn't matter in the grand scheme of things.
You could argue that trying to pull developers from Linux to stick with Windows, and just use WSL for ML/AI/CUDA is part of an extinguish strategy. It is when you think of it quite literally, as keeping people on WSL where they control it. All depends on your point of view and what we're referencing when we say extinguish.
But the landscape is different. EEE applied to things microsoft could control in some way. The battle fought against Linux was lost long ago and the foothold that Linux has in the total market is way too big for EEE to work. Sure they can do the first E. They can try to do the second E, but it is doomed to fail because it won't find significant adoption. Which means the third one is out of the question.
To me the whole WSL thing is a curiosity. They made a Linux kernel work on Windows in a VM like manner, but what is the point? To have your Linux service go down with a windows update? Making things dependent on two systems instead of one is a guaranteed way to make it less reliable. As a whole WSL actually looks the same as Stadia. Sure, there is a narrow niche that might see an advantage in it and would use it, but it is not objectively better and will therefore be ignored by the big crowds.
(I tend to think in a similar manner when talking about SteamMachines or Valve's Linux support: The goal of these two exercises is rather long term, and probably quite different that an immediate user adoption/market share/business revenue).
appetrosyan May 20, 2020
Patreon
PayPal
See more from me