Microsoft Becomes Linux Foundation Platinum Member
linuxfoundation.org
linuxfoundation.org
It's an important political gesture that Nadella goes in this direction. Since they also added a linux subsystem into the latest Windows release[4], I get the impression that he wants to leave the cloud to linux and try to position Windows as a user-facing client. This is a difficult decision to make, but it makes sense. Microsoft without Ballmer is seeing its position in the Corporate world as it is and I hope we will continue to see more openness as a result.
[1] http://www.slashgear.com/hp-pays-500000-for-linux-foundation...
[2] https://www.microsoft.com/en-us/Investor/earnings/FY-2016-Q4...
[3] https://en.wikipedia.org/wiki/Embrace,_extend_and_extinguish
It makes financial sense for them to join the Linux Foundation. Joining the Linux Foundation does not make them to, for example, start contributing to WINE, since that does not make financial sense to them.
HP's donation isn't necessary Microsoft's
> so it's not that big a deal for Microsoft
So what? It's a big deal for the Linux Foundation. And that's what matters. It's not about everyone paying "their fair share". It's about ensuring a good future for the high-quality very popular open source OS Linux. The circumstances of the donor doesn't affect the efficacy of their donation.
I hope we focus more on the outcome than the drama.
If they weren't, they would have never pushed hostile proprietary vendor lock in tech like DirectX12 or Universal Windows Platform in recent years.
"Linux was originally developed as a free operating system for personal computers based on the Intel x86 architecture"
"Linux has the largest installed base of all general-purpose operating systems."
"Linux is not the only such operating system"
"Linux is also the leading operating system on servers and other big iron systems"
I'm surprised you didn't distinguish it as an OS kernel, instead of leaving it ambiguous as to what kind of kernel it is, e.g. popcorn.
(2) The opposite is true. For example, Ubuntu user-space (i.e. sans Linux kernel) is available on Windows.
The operating system's technical name is GNU/Linux[1]. Many people call it Linux incorrectly, but it is actually a serious distinction[2]. I would agree however that Linux is the most widely used OS kernel.
To make the distinction more obvious, can you explain what the difference is between "Linux" and Android (which uses the Linux kernel). [ Hint: the difference is that Android doesn't contain GNU. ]
[1]: https://www.gnu.org/gnu/linux-and-gnu.en.html [2]: https://en.wikipedia.org/wiki/GNU/Linux_naming_controversy
Almost no Operating System is precisely GNU/Linux - most have vim, an X server, GNOME or KDE, 7zip etc. And the argument that GNU is a more principal argument is incorrect.
I don't considered everything licensed under the GNU license, or formerly associated with the GNU project / FSF to be 'GNU'.
[1] https://kernelnewbies.org/Linux_3.3#head-b733d694037e0b34ad4...
The "operating system" is not "technically" "GNU/Linux" - that's just what Richard Stallman wants it to be.
Stallman wants the definition of "operating system" to be the fully usable system. In which case Linux won't work. You need GNU coreutils. But you need a lot of other things too, why are the other things omitted from the naming?
Windows on the other hand is clearly an "operating system." Does this mean that the OS is "Ubuntu" or "Alpine"? Or are they just distributions still? In this case GNU/Linux is still not the OS.
Dennis Ritchie once referred to the kernel as "the operating system proper", albeit when describing various parts of unix (https://www.youtube.com/watch?v=JoVQTPbD6UY#t=1m28s). I think Ritchie's opinion is much more valid than Stallman's. [edit: this doesn't prove that the operating system is just the kernel, but that it's not far fetched to refer to the kernel as an operating system, or in that case to refer to an operating system by the kernel]
The short of it is that when people say that Linux is an operating system, perhaps they mean "a family of operating systems based on the same kernel" but that's a mouthful. And since most people would include things like Alpine in there when they say "linux", they clearly arent just referring to GNU/Linux. Thats one of those things where "Windows 10" is a different version of "Windows" than "Windows 8" because the bundled user space is different... but theyre both "Windows". If something requires Windows 10, I can say "requires windows 10". If it just generically works on windows 3.11 until present, I can say "works on windwos". "Alpine" is a version of "Linux", and so is "Ubuntu". Ubuntu just also happens to fall under "GNU/Linux". If something requires GNU, i can say "requires Linux with GNU" and otherwise "requires Linux" is absolutely correct. If something requires a specific distro, then you can specify that, otherwise you go for the lowest common denominator. (a lot of things that "require linux" will work on unix/BSD too, but no one throws a fit over that).
Also, the most important factor is that communication is about conveying thoughts effectively. If you understood "Linux" well enough to try to correct it, then you understood and thoughts are conveyed. That also means insisting on GNU/Linux for colloquial use is just pedantic.
If someone says just "Linux" they colloquially mean "Linux based operating systems" or perhaps even "Linux based operating systems like those in the most common distributions".
If I say "IE only runs on Windows", people dont argue that it doesn't run on Windows 3.11. They just know that "Windows" means "modern Windows" in this case, or at the very least, the implication that IE DOESNT run on competing operating systems. I mean, if I'm in a debate about Microsft vs Apple and I say "but for DirectX 12 and games you need Windows!" - here I specifically mean Windows 10, but "Windows" is enough to distinguish it from competitors.
If I say a game is for "Windows and Linux" or "Windows and Mac, but not Linux", its safe to say you understood what I'm talking about. And arguing otherwise means you're being pedantic.
If we're not currently in a discussion about differences between kernels or even libc implementations (in those discussions, it's safe to assume we're not using a colloquial definition), then it is safe to say that "Linux" means "Linux based operating system"
I'd have no problem calling Alpine musl/Linux or whatever.
> You need GNU coreutils. But you need a lot of other things too, why are the other things omitted from the naming?
Calling GNU/Linux also reflects the history of the operating system. This is actually the main reason that Stallman objects to calling it "Linux" -- because you're ignoring the fact that GNU came first (and the reason why GNU exists, which lead to all of the other projects you're talking about). You wouldn't have X.org, vim or many other projects without GNU -- simply because the community wouldn't have existed.
> If something requires GNU, i can say "requires Linux with GNU"
... no. Because you can run GNU/kFreeBSD, GNU/kNetBSD or even GNU/NT. How would you describe GNU/kFreeBSD to someone (remembering that to users it is basically indistinguishable from GNU/Linux)? "It's like Linux but doesn't have Linux in it?"
Exactly like that. I've heard people refer to all posix systems blindly as "unix" and its not worth correcting. If someone knows what GNU/kFreeBSD is, or cares to discuss it, we can discuss it. But generally speaking, I try to keep things concise. When I say "linux" im not trying to go into details, otherwise I'd specify that I'm using arch, ubuntu or whatever. I'm trying to lay down a generic concept of a group of operating systems. And saying "GNU/Linux" doesnt help me convey that idea any better to 99% of people. "linux" is just slang for "gnu/linux".
So if you want to say gnu/linux out of respect for stallman, sure. go ahead. but im not going to.
Just because companies release competing, incompatible products, doesn't mean they're "hostile" towards one another.
If you honestly think Samsung would willingly hand over $10-15 per handset over bunk patents, I've got some oceanfront property in Arizona you might be interested in.
That has nothing to do with Linux or Android directly, and everything to do with ANY CELLPHONE.
These patents aren't being used passively. They are being used for royalties. And Samsung wouldn't pay the royalty unless there was a strong legal case.
Microsoft could push Windows all they want if they were not also pushing an ecosystem exclusive to that platform along with it. If Windows was "just another unix" and was interchangeable at the user level with other operating systems, particularly those that would respect user freedoms, there would be no issue with its nature as "default" or in Microsofts behavior in regards to its users, because there would be real choice.
Real choice, in exactly the same way Red Hat vs SUSE vs Ubuntu users have right now amongst themselves.
What's my choice if I'm an AIX user, or a Solaris user, or an HPUX user, or a Windows user, or a FreeBSD user? My choice is to switch operating systems? Sort of like the choice MS provides?
In my opinion, they are not at all interested in cooperation until they port Office and Visual Studio to Linux or open source them.
But I'd argue that the most used microsoft product in the world (besides Windows) is MS Office. What holds a lot of people back from running Linux is that they don't like dealing with compatibility issues with LibreOffice and MS Office.
Microsoft knows if they open source or port MS Office to Linux, they will lose a massive market share and millions of people will stop using Windows. And I don't blame them, any reasonable business would not do something like that.
But unless they do that, I don't see how they can claim to be cooperating when they refuse to port their main product to Linux.
That's a huge assumption. There's no real indication that people would stop using Windows if Office existed in Linux.
If that were the case, then porting Office to OS X would've had a similar effect, don't you think?
Ultimately I believe the reason Microsoft doesn't bother porting to Linux is that the market-share is just not worth the cost of re-writing for another platform, and keeping it updated.
Additionally, as a compromise Office 365 runs on browser, which further decreases the incentive to port to Linux. Since you can get a similar experience using Office 365 on Linux, that said I won't be so disingenuous as to imply that the experience is as good as a native application.
In conclusion, Linux has a very small market share, so making a native application for it might not be cost effective. Additionally, Microsoft might see that market share as "served" regardless due to their online Office 365 web application.
Except for basically every thread I've seen about switching to Linux, where people say the only reason they won't switch is because they still want to use Office and Photoshop.
>If that were the case, then porting Office to OS X would've had a similar effect, don't you think?
No. Cost is a major reason people don't switch to OS X. But Linux is free.
> Ultimately I believe the reason Microsoft doesn't bother porting to Linux is that the market-share is just not worth the cost of re-writing for another platform, and keeping it updated.
That is a good point. But as I said, unless they port their main product to Linux, how can they claim to be cooperating when the ONE product that most people agree is a very good MS product is not available natively on Linux?
Cost is a lot more than the physical machine.
This is why Linux machines while technically less expensive (no OS license) are actually more expensive in terms of support. Until that nut can be cracked it's hard to get wide-scale adoption (e.g. 50% deskops vs. 5%).
It really depends on the company and the nature of work people are doing. Some software is a mess on anything but Windows. Some software is actually better on Linux.
Those are people who are aware of Linux, while much of the population isn't. Selection bias at play.
> unless they port their main product to Linux
Office accounts for about 5% of MS's revenue.
> how can they claim to be cooperating when the ONE product that most people agree is a very good MS product is not available natively on Linux?
That's not the only definition of "cooperating"...
How do you think the internet was made? By setting aside proprietary protocols and agreeing on open standards. Each vendor's free to do whatever they want so long as they use those standards.
Is that why they're pushing the Azure cloud service - to benefit Linux?
The important thing about this news is not the title or the amount. The important thing to notice that Microsoft has moved from extreme hostility to active co-operation with general Linux community. This is a good move and it benefits Linux community lot more than Microsoft.
I always felt that it was needless to paint MS/Linux as some kind of zero some game to begin with. Sad that MS's top leadership fell prey to it in the initial years trying to hurt Linux.
I think they are doing this because they think they have to and not because they have actually embraced anything. I think they are trying to buy goodwill.
NT will be with us for a while yet, though I can't understand the business case for not end of lifing it at some point, especially when most of the engineers you hire don't know it well.
All of the casual users that don't care about open source will just give in and use Windows as their OS when they hit a driver snag or something since they still get the userspace tools they like.
If anything Microsoft is providing a way to encourage Windows users that would never touch a BSD or Linux userspace tool to give it a try. That's a possible way to bring new users to Linux: teach them how to do things with Linux userspace tools and encourage them to dive deeper until they can do everything they want with Linux userspace tools and then maybe encourage them to switch.
If anything, Microsoft has given the Linux community the keys to possibly embrace some "casual users" that otherwise never would have explored Linux.
We can't forget about the bad MS is still doing: forced Win10 upgrades, Win10 spyware, Android patent extortion, FAT/VFAT/exFAT patent suits and licenses, etc.
The developer arm seems to be working hard, which is commendable. But looking at the rest of the org, this just seems like another Embrace round of the EEE cycle.
[0] https://en.wikipedia.org/wiki/List_of_Linux_kernel_names
This.
It doesn't take much skepticism to suspect MS has the ultimate agenda of eliminating all things Linux (as it exists today).
I'd welcome that, but MS still needs to fix several very major issues which stand in the way of such cooperation.
1. Stop the patent aggression.
2. More support for open standards. Especially in 3D graphics (DirectX lock-in), filesystems on removable media (exFAT lock-in), and so on and so forth.
3. Stop Windows tax / Windows bundling anti-competitive practices.
Once those are changed for the better, I'd say MS really changed. There is some progress with the above for example in case of the browser. MS joining Alliance for Open Media is one such case. But quite a lot still remains problematic.
well bundling is not a problem for me. I mean most people will use windows no matter how much better linux gets.
however the bigger problem is that windows actually allows vendors to create shitty drivers and fix shitty hardware within there kernel, that should not happen. windows should ONLY allow signed drivers for companies that are willing to create specifications for their devices.
4. What really irks me, is a point you missed: they are dealing with corrupt governments everywhere to make sure all government and education computers have Windows and Office, and that is what gets taught in schools.
In many countries yes, they are still doing it.
Microsoft was late to the game, but we gotta give them props for contributing back something more useful than "open core", platform onboarding or "hire me" piles of code.
I'm fairly certain all of 2016 is a mass-hallucination. Or something. In all seriousness though, I don't think this is surprising. Visual Studio on macOS is more surprising to me. Azure runs on Linux and that's a really, really big business for Microsoft. And they've also built a Linux Subsystem into Windows. Rock on Microsoft, rock on.
Edit: clarifying
When you're running Linux on Azure, it's via Hyper-V or rather the Azure flavored form of it. Windows is very much the core of Azure.
Azure Cloud Switch does use Linux - on metal I believe so no virtualization.
Edit: a word
My source: I work for MSFT, not on Azure but I work with their team a lot and I used to work on Windows.
I am happy to use Office in wine, as it gives me a very stable experience, full unicode supports even with my Xorg keymaps.
I use Visual Studio in wine - not as good as Office, but still very good.
I have a Windows phone, a priced relic since Microsoft abandoned them. Rock stable, last for a week in airplane mode.
I have a Windows 10 LTSB in a separate partition. Initially I just wanted to test it out. Now I consider running it in KVM for Visual Studio, in case I want more than what wine can now offer.
Hell, I am considering replacing my Thinkpad by one of Microsoft surfaces. As soon as it runs Linux as well as a Thinkpad, can get 32G of RAM, a user replacable SSD, wifi and lte module, I buy one. Seriously, even if the keyboard is not as good. And given Microsoft new focus on developpers, I wouldn't entirely dismiss the idea of a Surface Developper, bulkier but sturdier and user upgradable.
People say "hell froze over". I don't care. I look forward for more change from Microsoft. Because their tools are innovative again.
It'd be great if companies would support this sort of use if they're not going to bother creating native versions.
I used to play games using PlayOnLinux but the company, Origin (IIRC), updated their game manager software seemingly in order to break the Linux compatibility. The games were paid for and worked perfectly well.
Annoying as hell when companies appear to be actively hostile to you buying their product.
/me ducks.
'dang, go ahead and do you thing.
I guess the money from the LF is helping pay Linus' paycheque, but it has also become a bit of a lobbying group to steer Linux into more proprietary software.
Let's see how they'll try the next step.
Once open source, it's much harder to extinguish.
You have to think more business like. Why extinguish Linux when every copy running somewhere can give you licencing income. Maybe they'll partner with redhat and provide them something that has to be paid. Unlimited possibilities really.
I believe Microsoft is doing these moves, because they have decided that the old business model where all products are required to support the sales of each other is too risky. It served them well, but now there are too many good alternatives on many areas. If Azure does not support Linux, customers are not going to migrate to Windows because of that, they just pick another cloud. The old way was risky, because it meant that if customers moved away from Windows, they also had to move away from .NET and SQL Server.
Sorry if I said something silly :)
So not much, unless the work they have to do is highly parallelizable (i.e. can be broken up), and would truly benefit from more developers.
I can imagine KDE benefiting from this, because KDE is an umbrella organization with hundreds (if not thousands) of projects underneath it.
Just a really easy example is how they now have the Kirigami UI framework, and they released their software store redone in it a few months back (Discover) but the man hours and work required to port more of the collection to that toolkit would be immense. And even Discover is missing a lot of polish since its primarily a demo product at this point.
Throw another 10 developers in there (at 50k a pop?), and all those complaints about "it looks like Windows" with toolbars and dropdowns would be abated.
A better funded Linux Foundation could also invest in better tooling (better compilers, better test rigs), and in research on speculative improvements to pretty much every subsystem.
But on that note, I never understood why they track the drivers in the same git repo as the kernel.
They should just develop a stable kernel API for drivers, and have extract the drivers out and track them separately from kernel development.
Because the advantage of Linux drivers is code sharing, cooperation between different companies include competeting one, hierarchy of maintainers and developers. No driver code can be merged into upstream driver until subsystem maintainer, Linus and planty other developers see it.
So companies that maintain drivers in same subsystem improve and fix each others code. They also can't just merge some mess into upstream and have to follow certain rules that benefit kernel as whole. E.g for instance at least in GPU drivers it's not allowed to add any code upstream if it's only used by proprietary components in userspace.
> They should just develop a stable kernel API for drivers, and have extract the drivers out and track them separately from kernel development.
https://www.kernel.org/doc/Documentation/stable_api_nonsense...
i think the frontier you are speculating about is currently at 2 or 3 times that magnitude. what does a $100B-$1T of market capitalization behind linux and its ecosystem look like?
i think the major players are aligning to answer that question... we'll find out soon enough.
Open Source != Free Software
[0]: https://www.microsoft.com/en-us/Investor/earnings/FY-2016-Q4...
The company has become an enthusiastic supporter of Linux
Curious to know from any users out there: How well does Mono work, especially on Linux?They all work really well on Linux. They're part of the .NET foundation (https://github.com/dotnet).
Windows subsystem for Linux https://blogs.windows.com/buildingapps/2016/07/22/fun-with-t...
Xenix https://en.wikipedia.org/wiki/Xenix
Microsoft Interix, Windows services for Unix historic link https://en.wikipedia.org/wiki/Interix
(Social game: what will they call it? I'd say something like Unsinkable Linux ;-)
They should also have their own filesystem - ow, right, NTFS. I was thinking they could take another go at the database FS that stalled Vista a decade ago ...
http://www.infoworld.com/article/3042247/open-source-tools/b...
http://www.linuxjournal.com/article/5015
https://slashdot.org/story/04/07/11/1714235/gates-open-sourc...
It's a trap.
Been around long enough to have formed that singular impression of such Microsoft initiatives and to have it confirmed, repeatedly.
They failed in mobile. They are losing mindshare elsewhere.
It's a trap.
I'm not going to try to back up my statement, here and now. Just keep it in mind. The more wary you remain, the quicker you can move away when the noose tightens.
Cheers
Remember that the majority of people working at Microsoft today probably have little to no memory of the tech world of the early 2000s. They don't seem too much better or worse than others in attempting to shape open-source projects to their benefit - that's something you should always be wary about.
https://eugene.kaspersky.com/2016/11/10/thats-it-ive-had-eno...
Obviously Kaspersky isn't open source, but the tiger's stripes certainly haven't changed.
Also, Microsoft bundling Windows 10 upgrade malware into its security update channel has turned me off Microsoft, probably forever. Still don't understand why installing 10 after being told not to isn't a CFAA violation.
Windows Server has controllable updates and reboots. Win 10 - is a client OS and wasn't designed with the same level of hardening and shouldn't have the same uptime requirements.
The challenge is convincing the general public of this; people want a Ford Fiesta with the speed and features of a Audi R8, meaning they want the best of all worlds at reasonable cost and no penalties.
By forcing updates on clients, MS has eliminated a generation of machines being owned by various botnets. They've also made it relatively cheap to get a server in Azure should you have a temporary need for uptime (like a research project for a week).
That aside, some portions of those updates are bad (like removing functionality, pushing telemetry, etc), but the overall strategy of auto updating isn't a bad idea if they can convince people to adjust to it and use Server when they need stability.
Having a limit on the number of antivirus products a computer has makes sense, antivirus software often causes conflicts with other antivirus software, and a large amount of security software on a PC causes it to run very slowly, as duplicated scanning efforts take up a majority of drive performance on the slow 5400 rpm hard drives shipped with most consumer PCs today.
To be clear: Antivirus companies have acted like a plague on Windows computers for a long time. They use scare tactics to convince people to spend money on placebo-effect tools that rarely protect them from the type of malware they actually are likely to encounter.
One of the favored strategies is to bundle on new PCs with a short trial, and then tell people they won't be protected unless they subscribe. People aren't told that if they just remove the third party product that was preinstalled on their PC, they'll have perfectly serviceable antivirus.
That isn't to say Windows 10 doesn't have issues that still bother me, but let's just say that Kaspersky crying about them doesn't bother me much.
Will you help wine devs with some source code?
Please continue moving in this direction. There's still a ways to go.
Of course there's still the licensing situation of FAT32, ActiveSync, etc - it's hard to interoperate when the other party will only speak restricted protocols.
We should be very careful with Microsoft.
In Microsoft's case, many people believe they keep showing their true colours from time to time.
On a side note, I'm also displeased with microsoft's products I'm almost forced to use. Can you imagine that uninstallation of Visual Studio 2015 from SSD takes 20 minutes? And it doesn't even fully uninstall it. Windows updates are no better. Really makes you wonder if it's still 1998 at Microsoft.
> I have lost all trust in them, they have to gain it back and can only prove themselves with actions now
Question: what actual actions could they take that would cause you to change your mind about them?
But I not against Microsoft in particular. They are rich and powerful. It would be good, if they are really good guys now. I give them a chance to prove themselves. But i don't expect much.
>what actual actions could they take that would cause you to change your mind about them?
Maybe start making good products? What more can they do? And going back to "new blood" point again, do you think noone in Visual Studio devteam noticed that uninstallation taking 20 minutes? I'm 100% sure that there is guy who did, but he just cannot do anything. He have no power to change things. "It's working now somehow, so don't even think about touching it." And everything is like that at Microsoft.
And that whole telemetry stuff in win10. Every single option (telemetry/internet search/ads etc) is turned off, yet i open up wireshark and still see it calling home every time i type something in start menu. It's like they don't respect me and think that i'm stupid. Why it has to be this way?
"We are different now, but we are still the same." And nothing has changed. Pretty sad.
When I first joined Microsoft, it was prior to the launch of Win8, and I saw a lot of complacency throughout many parts of the company. I found it frustrating, because I came from an open-source background on the Java stack, and I wanted to see more innovation.
In just a few years, things have changed dramatically, and that is because the market has changed. As others have said on this thread, Windows is no longer the key to Microsoft's business success, it is the ecosystem of products that Microsoft sells.
That is why there have been all of these cultural changes, because the industry has shifted, and there are some really smart leaders at Microsoft (e.g. Scott Guthrie) who realize that we have to change to stay viable.
So, the terrible practices you once saw just don't even make sense anymore. Microsoft is being challenged in every market it operates in, which means that we're forced to innovate. This turns out to be better for everyone, including employees like me. (I came from an academic research background, and I'd never want to work somewhere that was happy with the status quo.)
And yes, the Visual Studio 2015 uninstall takes a really long time. Some installs also take forever. They've made a ton of improvements in the new preview VS 2017.
The reason this wasn't done previously was not that a few engineers couldn't change things (though sometimes that is true), but rather that it was a hard engineering problem. It's easy to think that a big company has infinite resources, but the reality is that you can't build a product with too big of a team (see: Mythical Man Month), so you have to choose what you focus on. If you were to ask most customers, they'd rather have new features than have a better installation experience. That is, until the install experience got in people's way.
I don't work on anything related to the Windows client, so I can't speak to the telemetry, but what you're describing does sound pretty annoying. It is all anonymized (I know this because I'm familiar with the privacy reviews that all code has to go through), but I can see the case for wanting to opt out completely.
Obviously, one post is not going to change your mind (and I'm not looking to do so), but I'd suggest you look at leadership and philosophy changes, not just individuals.
That's what worries me the most. The "cool" guys won't stay there forver. You can never now who will be the next.
Go visit the VS2017 thread. You'll see that install/uninstall is much faster now.
Still, it'll be a while before they are worth our trust. They fought dirty and hard.