ZDNet Argues Linux-Based Windows 'Makes Perfect Sense' (zdnet.com) 100
Last week open-source advocate Eric S. Raymond argued Microsoft was quietly switching over to a Linux kernel that emulates Windows.
"He's on to something," says ZDNet's contributing editor Steven J. Vaughan-Nichols:
I've long thought that Microsoft was considering migrating the Windows interface to running on the Linux kernel. Why...? [Y]ou can run standard Linux programs now on WSL2 without any trouble.
That's because Linux is well on its way to becoming a first-class citizen on the Windows desktop. Multiple Linux distros, starting with Ubuntu, Red Hat Fedora, and SUSE Linux Enterprise Desktop (SLED), now run smoothly on WSL2. That's because Microsoft has replaced its WSL1 translation layer, which converted Linux kernel calls into Windows calls, with WSL2. With WSL2 Microsoft's own Linux kernel is running on a thin version of the Hyper-V hypervisor. That's not all. With the recent Windows 10 Insider Preview build 20211, you can now access Linux file systems, such as ext4, from Windows File Manager and PowerShell. On top of that, Microsoft developers are making it easy to run Linux graphical applications on Windows...
[Raymond] also observed, correctly, that Microsoft no longer depends on Windows for its cash flow but on its Azure cloud offering. Which, by the way, is running more Linux instances than it is Windows Server instances. So, that being the case, why should Microsoft keep pouring money into the notoriously trouble-prone Windows kernel — over 50 serious bugs fixed in the last Patch Tuesday roundup — when it can use the free-as-in-beer Linux kernel? Good question. He thinks Microsoft can do the math and switch to Linux.
I think he's right. Besides his points, there are others. Microsoft already wants you to replace your existing PC-based software, like Office 2019, with software-as-a-service (SaaS) programs like Office 365. Microsoft also encourages you to move your voice, video, chat, and texting to Microsoft's Azure Communication Services even if you don't use Teams. With SaaS programs, Microsoft doesn't care what operating system you're running. They're still going to get paid whether you run Office 365 on Windows, a Chromebook, or, yes, Linux.
I see two possible paths ahead for Windows. First, there's Linux-based Windows. It simply makes financial sense. Or, the existing Windows desktop being replaced by the Windows Virtual Desktop or other Desktop-as-a-Service (DaaS) offerings.... Google chose to save money and increase security by using Linux as the basis for Chrome OS. This worked out really well for Google. It can for Microsoft with — let's take a blast from the past — and call it Lindows as well.
That's because Linux is well on its way to becoming a first-class citizen on the Windows desktop. Multiple Linux distros, starting with Ubuntu, Red Hat Fedora, and SUSE Linux Enterprise Desktop (SLED), now run smoothly on WSL2. That's because Microsoft has replaced its WSL1 translation layer, which converted Linux kernel calls into Windows calls, with WSL2. With WSL2 Microsoft's own Linux kernel is running on a thin version of the Hyper-V hypervisor. That's not all. With the recent Windows 10 Insider Preview build 20211, you can now access Linux file systems, such as ext4, from Windows File Manager and PowerShell. On top of that, Microsoft developers are making it easy to run Linux graphical applications on Windows...
[Raymond] also observed, correctly, that Microsoft no longer depends on Windows for its cash flow but on its Azure cloud offering. Which, by the way, is running more Linux instances than it is Windows Server instances. So, that being the case, why should Microsoft keep pouring money into the notoriously trouble-prone Windows kernel — over 50 serious bugs fixed in the last Patch Tuesday roundup — when it can use the free-as-in-beer Linux kernel? Good question. He thinks Microsoft can do the math and switch to Linux.
I think he's right. Besides his points, there are others. Microsoft already wants you to replace your existing PC-based software, like Office 2019, with software-as-a-service (SaaS) programs like Office 365. Microsoft also encourages you to move your voice, video, chat, and texting to Microsoft's Azure Communication Services even if you don't use Teams. With SaaS programs, Microsoft doesn't care what operating system you're running. They're still going to get paid whether you run Office 365 on Windows, a Chromebook, or, yes, Linux.
I see two possible paths ahead for Windows. First, there's Linux-based Windows. It simply makes financial sense. Or, the existing Windows desktop being replaced by the Windows Virtual Desktop or other Desktop-as-a-Service (DaaS) offerings.... Google chose to save money and increase security by using Linux as the basis for Chrome OS. This worked out really well for Google. It can for Microsoft with — let's take a blast from the past — and call it Lindows as well.
Oh no (Score:3)
Dave Cutler must be pissed
Re: (Score:2)
Dave Cutler must be pissed
Dave Cutler was pissed pretty much from NT4 onwards, when drivers were moved into kernel space.
Now today, an errant driver can be unloaded and reloaded at whim (Windows - Ctrl - Shift - B restarts the display driver, for instance), but back in the NT4 an exception in a driver brought down the system. Unlike NT3.1.
If you're wondering where Microsoft makes money... (Score:5, Informative)
From their 2019 annual report. https://www.microsoft.com/investor/reports/ar19/index.html [microsoft.com]
Year Ended June 30, 2019 (In millions)
Server products and cloud services $32,622
Office products and cloud services $31,769
Windows $20,395
Gaming $11,386
Search advertising $7,628
LinkedIn $6,754
Enterprise Services $6,124
Devices $6,095
Other $3,070
Total $ 125,843
Re: (Score:2)
Re: (Score:3, Insightful)
Only idiots and companies interested in b2b sales and recruitment.
IIRC At the time of the acquisition LinkedIn was the only social network generating meaningful membership and service fees alongside advertising.
Re: (Score:2)
Are there that many idiots and companies buying whatever it is that LinkedIn in sells ?
It isn't "that many," it is that rich people get charged a lot of money for rich people services.
The only alternatives is to hire all your own staff, (but then you can't yell at them, or they'll quit) or to pay really high rates that poor people can't afford. And then whoever you paid will make sure that they have people you can call up and yell at, 24/7.
Re: (Score:2)
Targeted advertising (ugh).
Re:If you're wondering where Microsoft makes money (Score:4, Insightful)
Bill Gates' Microsoft Strategy: "Own the desktop!"
Steve Ballmer's Microsoft Strategy: "Own the user!"
Satya Nardella's Microsoft Strategy: "Own Linux!"
Re:If you're wondering where Microsoft makes money (Score:4, Funny)
Next Microsoft CEO: "We got 0wn3d!"
Re: (Score:2)
Not really, if you consider that 20 years ago, almost 100% of "Server products and cloud services" and "Office products and cloud services" would have been bundled as part of their corporate Windows supply contracts.
If you add those three categories together, that is what they used to make only because of their Windows monopoly. But they've already successfully migrated to selling those other services directly, on their own merits, in a market without platform lock.
There is really no reason for them to try
Re:If you're wondering where Microsoft makes money (Score:5, Insightful)
Windows might not be its biggest revenue stream anymore but you dont just walk away from 20 billion per year revenue with windows.
Re: (Score:3)
they will still sell windows, no other distro will have the windows layer!
but they will save lot of money by not having their own kernel, drivers and all the linux performance and features. Of course they will not use GNU libc, nor most of the GNU tools, they will use their own and block other from simply selling the same thing. Think MS doing what google did for android, the kernel is linux, but the remaining apps aren't
Re: (Score:2)
Anyway, free and open source Windows compatible layers already exist for Linux. It's called WINE. I see no reason Microsoft should compete with WINE, which would involve 10's of millions of dollars worth of development time and engineering to even come close to feature parity with the exis
Re: (Score:2)
Re: (Score:2)
They've already started on the basics of this though - take a look at how SQL Server on Linux runs, there's an emulation/API hook layer in there that handles the conversions.
But SQL Server is their own product, they did just what is necesssary and it's targetted for servers, not desktop. Windows is "known" for it's backward compatibility. And then think about their current drivers.
If they do a NEW OS to sell alongside Windows, then yes, linux kernel may be targetted.
But for replacing the current Windows kernel and keep all current apps&drivers working (otherwise who would use it), I would rather think they would go for a Unix kernel, not linux, if only because of licensing.
Re: (Score:2)
No need to port drivers, linux drivers already do that job. The few that may need, may be pushed to support linux or add BOTH NT and a new linux api, so can work on both
For apps, all they need is to support the documented API in their layer, after that almost all apps and DLLs will work. That is why they choose that path for the SQL server, not port the software to linux, but add a compatibility layer, so software could work with the minimum changes
Re: (Score:2)
Like other said, check the SQL server example and remember that today, most (MS) windows app are .NET, that can also be run in linux and that MS is expanding the .NET support in linux
Re:If you're wondering where Microsoft makes money (Score:4, Informative)
True, but the article isn't talking about them walking away from Windows revenue. It's about cutting development costs by using the Linux kernel instead of the Windows kernel. They would likely still provide a 100% Windows user space on top of kernel land. It would be similar to the relationship between Linux and Android.
It's an interesting idea, but it's easy to foresee a lot of pitfalls for Microsoft. Windows has long taken backwards compatibility to the extreme, and Linux doesn't have ABI stability. There are a _ton_ of cheap devices out there that only have Windows drivers. How would you get these to work in such a scenario? Could Microsoft make suitable modifications to the Linux kernel to permit Windows device drivers (wouldn't that be interesting?)
Personally, I don't see MS taking this route anytime soon -- but it is somewhat interesting to contemplate.
Yaz
Re:If you're wondering where Microsoft makes money (Score:5, Informative)
NDISwrapper has been around for ages. It has a few problems with Windows drivers liking lots of stack space, but other than that it is pretty good.
A patch to the Linux kernel to reserve more kernel stack space would solve most of the remaining issues at a modest memory cost, at least on 64-bit systems. 32-bit Windows 10 is dying anyway.
Re: (Score:2)
Re: (Score:1)
I can't help but remember years ago when Gates was quoted as saying that NT was "just another unix"
Re: (Score:2)
And that 7 billion in search advertising (who knew?) is nothing to sneeze at either. What do you think that would be without Windows defaulting your searches to Bing?
Re: (Score:2)
Windows might not be its biggest revenue stream anymore but you dont just walk away from 20 billion per year revenue with windows.
You don't walk away from it right now. If the revenue has dropped you would strongly consider alternative long term solutions if you spend anything near 20,395 in maintenance and feature "improvements".
One alternative may be to follow Apple's lead and look at BSD or Linux for the kernel and proprietary for everything user facing.
Hyper-V is moving to Linux (Score:4, Interesting)
This story fits almost perfectly in the story a few weeks ago on ZDnet to make linux a root partition for Hyper-V. (https://www.zdnet.com/article/microsoft-these-patches-aim-to-make-linux-run-as-root-partition-on-hyper-v/)
- With linux booting Hyper-V, Windows could then run as a guest OS while it's completely transparent to the user which OS is running which application.
- Switching the core OS to linux whilst maintaining compatibility would unlock a whole new world for Microsoft, which is finding itself more and more confined in their proprietary environment.
- This proprietary nature of Windows used to be an advantage, but with 100% of the competition being on some form of unix, it no longer makes sense.
- Ironically, Microsoft virtualising Windows on top of Linux would be identical to what Apple's done 20 years ago with the introduction of Mac OS X.
Re:Hyper-V is moving to Linux (Score:5, Interesting)
The problem with this idea is that it won't help in the way people seem to think it will. TFA mentions all the bugs they patched, but they were not Windows Kernel bugs, they were almost all user land components, drivers and some of the bundled apps. So switching to Linux would make zero difference on that front.
In fact switching to a Linux kernel would probably just make things far worse. Windows has a decent security model and it won't be trivial to migrate to a different one. Along the way there are bound to be new bugs.
They would need to do a lot of work on the Linux kernel to bring it up to feature parity with Windows too. For example Linux still doesn't support self encrypting drives very well, so Microsoft would have to spend time fixing all the issues that are blocking that and then port BitLocker over to work on it, just to get back to where they started from.
It made sense for Apple to do it because MacOS prior to X was in a pretty poor shape, and because backwards compatibility was not such a big issue for them.
Re:Hyper-V is moving to Linux (Score:5, Informative)
Windows has a decent security model
Everything else you've stated is true, except for this. Windows has the core design for implementing a decent security model but as supported in production, the existing security model is severely lacking. It's still a basic per-user access model, with a half-arsed not-quite-mandatory access control scheme slapped on.. with occasional clever exceptions (e.g. removable access tokens) to implement things like UAC.
By comparison, macOS has GUI-level isolation between unprivileged applications but lacks good ways to isolate everything else. Linux, as used on servers, isolates non-GUI applications perfectly with multiple security layers (seccomp, namespaces and SELinux/AppArmor) which all compliment one another. Windows has the potential to run services in a very locked down way but half-arses it. macOS basically got there overnight with Seatbelt, mandating that people actually use it for App Store software (or GTFO). They've also added other Mandatory Access Control layers since then to protect important user data even for apps which don't have their own independent policies. Meanwhile, Windows has Controlled Folder Access which is a complete pain in the ballsack, doesn't differentiate between data types and again only protects against unauthorised writes - nothing stops an attacker stealing your nudes.
Web browsers, UWP apps, Adobe Acrobat and Microsoft Office do make use of GUI-level isolation but even then, it's very limited. With Office and Acrobat you can't even edit documents, while with web browsers, all bets are off the moment 3D graphics is involved.
UWP apps make good use of AppContainer, which is awesome, but without traditional Win32 apps doing the same, what's the point?
Service hardening tokens limit the network damage and file write capabilities of Windows system services, yet Microsoft doesn't consider the WSH sandbox as a security boundary (just like UAC). Why? It's broken (only limits writes last I checked) and doesn't even support the lowest common denominator server roles properly (e.g. a web server), only services where all access is pre-determined and immutable. A better solution would have been to run all services under separate UIDs in the first place, not grouping them all together into SYSTEM, Local Service and Network Service UIDs. Linux gets this right by using separate users for every service which doesn't need to impersonate, then using SELinux/AppArmor which completely support mitigating access to resources in both scenarios.
Moving from Windows to Linux would be undeniably bad to do right now for a lot of reasons but security probably isn't one of them.
Re: (Score:2)
The problem with this idea is that it won't help in the way people seem to think it will. TFA mentions all the bugs they patched, but they were not Windows Kernel bugs, they were almost all user land components, drivers and some of the bundled apps. So switching to Linux would make zero difference on that front.
I think the difference will be astronomical since Windows doesn't have SELinux. Bugs in things like AD could be contained. That's be great for Windows admins. It would be great for Linux admins too who have to use spamd to filter mail from compromised machines. Think of the CO2 benefit alone.
In fact switching to a Linux kernel would probably just make things far worse. Windows has a decent security model and it won't be trivial to migrate to a different one. Along the way there are bound to be new bugs.
They would need to do a lot of work on the Linux kernel to bring it up to feature parity with Windows too. For example Linux still doesn't support self encrypting drives very well, so Microsoft would have to spend time fixing all the issues that are blocking that and then port BitLocker over to work on it, just to get back to where they started from.
It made sense for Apple to do it because MacOS prior to X was in a pretty poor shape, and because backwards compatibility was not such a big issue for them.
Sorry I don't see the difference. Has MS ever cared about backward compatibility? Their whole model is based around buy YYYY+1 version of MS Office or you'll be left behind.
Re: (Score:2)
Sorry I don't see the difference. Has MS ever cared about backward compatibility? Their whole model is based around buy YYYY+1 version of MS Office or you'll be left behind.
Well it's much better than with Gnu/Linux. And yes, they do (or at least did) make extra effort to keep older apps working. At least the popular ones. Games released for Windows XP (15 years old) still run great (with new drivers), if you solve copy protection issues. You take 5-year old source code for Linux (forget about binaries), and it won't even compile. Drivers or apps.
Re: (Score:2)
This story fits almost perfectly in the story a few weeks ago on ZDnet to make linux a root partition for Hyper-V.
So? The whole idea is ludicrous... Microsoft isn't ever going to release the source code for Windows under GPL (which is basically what they'd have to do to make this thing work).
Re: (Score:2)
do not mix topics, this is about MS replacing the NT kernek by the linux kernel, nowhere they said that they will release things in GPL. They will probably not even use ANY GNU tools or lib.
This is basically what google did with android, the base kernel is linux, but toolset is android... while most of it is open source, for MS, all or almost all will be still closed source. their apps will use the same DLLs, but those DLLs will call linux API or a windows compatibility API in linux (that can even be a GPL
Linux is for devices and servers, not desktops (Score:3)
Microsoft's goal is to keep developers and users on Windows. Microsoft is making sure the *nix toolchain works under Windows so developers stay on Windows. Much like macOS offering a *nix toolchain via their BSD environment made it an attractive alternative to Linux. With macOS and Windows+WSL2 offering the *nix toolchain various developers and users get the best of both worlds. The commercial software ecosystems and the *nix toolchain.
T
Re: (Score:2)
Agreed. People are getting confused by the additional Linux compatibility offered on Windows, reading more into it than really exists. If you look more critically at the issues, replacing the NT kernel really makes little sense.
No one here on Slashdot wants to hear this, but from a design standpoint, the NT kernel has several advantages over the Linux kernel. The NT kernel has a stable driver API, and communicates to mid-level OS components through a flexible translation layer, which is what enables the
Makes more sense than realized (Score:5, Interesting)
And what's more is that the money "stack" is moving further and further away from running a kernel or OS at all. Apple doesn't make money off iOS without monopolizing the app store and services, not to mention having the standard browser, witness all the lawsuits. But if the OS itself doesn't make money, why bother? Facebook and Tencent (China) are already trying to supplant the money making part of an OS with their own layer on top. News, payment services, games, whatever, they don't even have to pay for basic OS stuff. Microsoft has the exact same idea with it's "MetaOS" plans, sync everything across OS's, all Microsoft all the time, sounds like money to them. So who needs stuff money down the pit of maintaining your own kernel?
Re: (Score:2)
They're two very different models. Similar in services perhaps, but not at OS level.
Re: Makes more sense than realized (Score:2)
There is a reason MS has aggressively pushed Surface and other hardware devices.
systemd was a warning (Score:4, Funny)
Re: (Score:2)
All hail the king (Score:2)
Linux-Based Windows (Score:2, Insightful)
Really in what Reality?
Re: (Score:3)
NT is clone of VMS, not Linux, and VMS was not Unix based.
Re: Linux-Based Windows (Score:1)
Re: (Score:1)
Re: (Score:2)
There is nothing *nix about the WinNT kernel
Just beware of how it will be perverted. (Score:1, Troll)
How somehow, MS's weird "Linux" will become "standard" and force software developers to adhere to it, to not get complaints from users about it not working. ... those a
And then your favorite Linux distribution will be forced to add those weird un-Unixy things too, that clearly come from somebody who does not get why things have been designed that way, and does not care either. Because otherwise, applications will have trouble running on your distribution, and people will complain. Clueless people, of course
Re:Just beware of how it will be perverted. (Score:4, Insightful)
And then your favorite Linux distribution will be forced to add those weird un-Unixy things too, that clearly come from somebody who does not get why things have been designed that way, and does not care either.
Lennart Poettering?
Other way around (Score:1)
Re: (Score:2)
I tend to agree. MS has hacks in the NT kernel to make their apps run decently even if their UIs are crap. They won't get that with Linux unless they fork Linux, and then it isn't Linux. Of course they'd probably still call it Linux.
Re: (Score:3)
The countless manhours wasted on WSL1 have given them insights on how to implement Linux system calls in WinNT.
If they ever implement LSW, mapping WinNT syscalls to Linux is a known quantity in Wine.
Xenix (Score:2)
Office online has a long way to go (Score:4, Insightful)
A load of utter BS (Score:4, Interesting)
There's no way the Linux kernel will ever be able to replace ntoskrnl.exe for the following reasons:
* The Linux kernel has an unstable absolutely different API for pretty much all devices. This transition if it to ever occur, would render thousands devices inoperable, this would also mean Microsoft will have to rewrite basic OS components like filesystem drivers. They will also have to rewrite tons of userspace code. The expenses will not be massive, they will be outright insane.
* Unlike Linux, the Windows kernel and userspace are tightly connected and integrated to provide the highest performance and integration possible.
* Microsoft has not control over Linux kernel development and direction. This makes the development of new features quite complicated. Google has been in the same boat for many years and they're so fed up, they are now developing their own microkernel (Google: Google Zircon).
I'm deeply sorry, but the Linux kernel will never replace ntoskrnl.exe in Windows unless Microsoft decides to replace Windows with (its own) Linux distro + Wine on top of it.
Re: (Score:2)
Unlike Linux, the Windows kernel and userspace are tightly connected and integrated to provide the highest performance and integration possible.
lolololol
The question isn't whether you drank the kool-aid, it's how much
Re:A load of utter BS (Score:5, Interesting)
Actually, the Linux kernel supports older devices that MS Windows has dropped. Also, Linux uses 1 driver to support multiple devices especially when the protocols are open standards.
The Linux kernel and userland are connected via libraries where the translation to kernel system calls is performed such as glibc for C based programs. This flexibility allows Google to implement Android by replacing the libraries.
Microsoft is a member of the Linux Foundation and therefore does have a voice in the evolution of Linux. Microsoft also provides patches into the Linux kernel git tree.
What you are forgetting is that the Linux kernel is very popular on ARM based embedded systems. In other words, Microsoft is trying to get a foothold into the ARM based consumer products where MS has traditionally been weak in.
It could be argued that the market for x86 based systems is now the niche market and ARM has won except for the desktop but this market is in decline.
Re: (Score:2)
Actually, the Linux kernel supports older devices that MS Windows has dropped. Also, Linux uses 1 driver to support multiple devices especially when the protocols are open standards.
And, Linux doesn't support thousands of devices that only have a small, Windows-centric user base. And, because it is open source, many devices are supported by volunteers who can, and do, just stop supporting whatever driver when they stop using that particular device. Kind of like how there are now thousands of unsupported open source projects that were started by college students who now have real jobs and lives beyond writing code.
Re: A load of utter BS (Score:2)
You guys are getting way too excited about Windows running a Linux VM, it's sad and funny at the same time.
It's all about attracting developers and admins from Macs and generally gaining back pro-nerd mindshare they've bleed over the past twenty years. It's funny because it's a good idea and actually working, it's sad because WSL is a better Linux than Linux... like in practice, zero f's are given about Linux's drivers or desktop software. WSL doesn't even need any more Windows integration all they need i
Re: (Score:3)
the Linux kernel will never replace ntoskrnl.exe in Windows unless Microsoft decides to replace Windows with (its own) Linux distro + Wine on top of it.
And they could very well do that. Microsoft would build a "Wine-on-steroids" compatibility layer over the unstable, bubbling septic tank that is Windows 10, and 99% of users would never know the difference.
Hell, MS could boot directly into a Windows VM running on Linux and again, 99% of users wouldn't know the difference. And they wouldn't care.
Honestly, it seems that my Win 10 VMs run better under Linux than native Win 10 does installed on bare metal.
Re: (Score:3)
Hell, MS could boot directly into a Windows VM running on Linux and again, 99% of users wouldn't know the difference. And they wouldn't care.
You're right about that. Windows 10 isn't actually running bare metal by default on modern hardware. It runs inside of Hyper-V (alongside a picokernel) to provide Virtualization Based Security, Credential Guard and HVCI. Have you tried turning all the extra enabled-by-default security off and then compared the difference?
It's a definitely a night and day difference on older hardware (anything older than 8th Gen Intel processors at least) and probably explains why you see much better performance with you
Re: (Score:2)
I'm deeply sorry, but the Linux kernel will never replace ntoskrnl.exe in Windows unless Microsoft decides to replace Windows with (its own) Linux distro.
I suspect Microsoft will sooner open source ntoskrnl within the next 10-15 years. However, I do think they'll adopt Linux for parts of Azure in order to stay competitive with Amazon and Google. Otherwise they'd have no reason to contribute code intended to allow Linux to function as a Hyper-V host, since it's already a perfectly stellar guest (and so is FreeBSD, minus the false positive SCSI errors from time to time).
Google has been in the same boat for many years and they're so fed up, they are now developing their own microkernel (Google: Google Zircon).
As in Fuchsia? I hope it puts an end to Android's obsession with creating piles of e-wast
what value does it bring? (Score:2)
I don't see in there what value having Linux "somewhere in there" brings overall?
Is it faster? Is it more efficient?
Re: (Score:3)
It IS faster and more efficient.
But that's not the real benefit.
The real benefit is that lots of other people do your work for you.
Microsoft is getting worse and worse at doing their own work, so that seems like it would be appealing.
Look at how many times there's been Windows 10 updates that kill your install. They probably add up to all the times Windows 7 and Windows 8.1 have done that combined.
Linux desktop is over. Move on. (Score:3)
Software-as-a-service? (Score:5, Insightful)
Why not call it what it is: "User-is-a-sucker".
My copy of Office 2007 cost about $400 back in 2008 or so. It still works perfectly well. If Microsoft 365 had been available back then, so far it would have cost me $109 per year (the family plan) for 13 years, or about $1,400. And for that princely sum, I'd have had the joy of knowing that if for any reason, my fault or theirs, Microsoft decided they hadn't received their monthly tribute, I would have lost access to Word, PowerPoint, Publisher, Access, Outlook, OneNote and Excel, and perhaps some or all of the files I'd generated using them (if I'd stupidly opted for Cloud storage).
So my answer to the question, "Given a choice, will you ever allow 'software-as-a-service' on your computer?", should be pretty obvious.
In practical terms, the first time I was offered that kind of "deal" was when PhotoShop became a service. I switched to the GIMP. It's not quite as good, but it's more than adequate for my needs. So if Microsoft tries to force me into one of their "User As A Milk Cow" plans, LibreOffice here I come.
Force us all to SaaS (Score:2)
SaaS isn't doing so well this week, with Office365 having a couple outages.
However, companies like it because they get paid to maintain software instead of develop it. Let's be honest: all of the big types of software (office, photoshop, browser, editor, audio, etc.) are already here.
There's no big score anymore in inventing a new program and selling a gazillion copies.
Instead, these companies are making their money by selling newer versions of the same old stuff... but people aren't buying, because the new
Re: (Score:2)
I have some (non-MS) Windows graphics/GIS software that has lost me as a customer by doing exactly that. They no longer sell "perpetual licenses." Everything is an annual license with a small discount for a 3-year subscription. I currently have the next-to-last old version that's "perpetual" though no longer supported (it never really was; there was only one update, shortly after it was released, to fix a showstopper bug). Support is now only for the subscriptions. It's good software, and I'll keep using it
Adobe is brutal (Score:2)
At this point, I'm telling clients that it's worth the training costs to move to FOSS. Adobe wants to squeeze $500 per seat per year out of anyone who still uses their products. It's a dying industry, making its customers into the cash cows.
GPLv3? (Score:2)
Re: (Score:2)
Certainly they would have to do that. But they have a horde of lawyers that could probably work on that in their spare time.
It is a done deal already (Score:1)
Massive Data Centers die for a 2% improvement in scaleablility. It could be the Linux kernel running Windows OS is 4% faster and Linux in the Azure cloud runs 30% faster than on Hyper-V. If you think I am stupid just look at how insane they get with heat, cooling and open compute motherboards. On the Open Compute side if a microprocessor for the sound card takes 1watt of power and costs $0.10 they nix it.... We are talking at scale, 100,000 servers or more per datacenter and the desire to have 1,000 dat
They would never (Score:2)
In 1982, I was working for a manufacturing division for the Conrac corporation.
I was in the process of finding a replacement system for their first inventory computer system.
Although we went with a little VAX running UNIX SysIII (I was a full out zealot even back then),
an IBM sales guy visited to give his pitch and I asked him about running UNIX on their hardware.
He laughed and said "You don't understand, IBM will never sell UNIX".
If they had listened to me... (Score:3)
Re: (Score:2)
Back then in 1990s it would not have been Linux chosen, rather a flavor of Unix, either BSD or System V, or possibly VMS.
Re: (Score:2)
It only makes perfect sense if you're an idiot (Score:2)
Fait Acompli (Score:1)
There is *already* a shitload of Windows running on Linux. You cannot get a mainstream Linux distribution that does not pretend to "look like the Windows clickety-pokey".
Windows means the Windows Kernel, and Linux means the Linux Kernel. They do not refer to the clickety-pokey "application shells" running atop that kernel.
Microsoft is no more about to "replace" the "Windows Kernel" with the "Linux Kernel" than the pigs underlying the flying pigs (the pigs with wings) will be replaced with dogs. Just beca
Re: (Score:2)
You seem salty. My MATE desktop doesn't look anything like Windows. Windows has all kinds of quality issues, even to bricking people's machines with bad updates. There is no reason for Microsoft to keep that troublesome core garbage, it makes sense for them to go to a capable kernel that scales from embedded systems to supercomputers.
Re: (Score:1)
I haven't seen any "quality issues" in the Windows kernel for many ages. Yes, there are security holes and stupid mis-design features (about on par with Linux or any other Operating System of consequence you care to name). Almost all of the so-called issues have been with the clickety-pokey layer, not with the Operating System itself.
Re: (Score:2)
Plenty of kernel bug patches come out each year. I still see occasional blue screens among the dozens of Windows servers we have at work.
Re: (Score:2)
Re: (Score:2)
wrong, neither GNOME nor KDE have Windows "look and feel". You are ignorant of the history of GUI. Guess again, I'll give you a hint. I had nice desktop on Unix boxes circa 1990.
Re: (Score:2)
I had nice desktop on Unix boxes circa 1990.
That's nice. I had a Linux desktop in 1990. I set up X on it manually. I used TWM.
neither GNOME nor KDE have Windows "look and feel"
You have a very short memory. I remember when they were first released I built both of them from source. They BOTH used the Windows look and feel. They both had a task bar, used a form of the "Start" button, etc. They were Window look-a-likes. People like you were denying it even when the similarities were listed. Your ilk copied the Windows look but you couldn't deal with it so you deluded yourselves and made yourselves look
Re: (Score:2)
File Systems (Score:2)
> you can now access Linux file systems, such as ext4, from Windows File Manager and PowerShell
Finally? Linux has been able to fuse native and NT filesystems for ages.
Re: (Score:2)
You not liking it doesn't mean PowerShell isn't a good tool.
Sooo.... what you're saying is (Score:3)
THIS is FINALLY the year for Linux???
Whew! I was starting to give up on all those prior declarations of Linux finally breaking out!
Re: (Score:2)
THIS is FINALLY the year for Linux???
Nope, PSYCH!
The latest Tech News Zombie... (Score:3)
No, this is not going to happen. Microsoft would have to make way, way too many changes to Linux to even hope to get it to where it needs to be for desktop and interactive applications that use Win32 and so on.
For one, you need a stable and documented kernel ABI and that still hasn't happened after decades. There are scheduler changes to be made and a ton (I mean a ton) of work to get video and audio where it needs to be. And that's just to start supporting for WDM and UMDF drivers. That is if the Linux team will allow binary drivers at all.
Or, they spend all their time improving the Windows kernel making it better at what it does and leave Linux on it's own path.
What took then so long? (Score:2)
I dumped Windows in favor of Linux over 20 years ago. Did it really take MS this long to figure things out?
Of course, the downside of this change is that Microsoft will figure out a way to make us pay for their version of Linux/Windows.
If Windows went Linux-based... (Score:2)
then Microsoft would start swamping upstream Linux with their code.
Any time spent working in tech support especially with, say, the ability to use strace to see what their code does, teaches you that whether it's down to sausage-grinder programming conditions, management, the Microsoft APIs, whatever... it teaches you that their ability to write good clean code SUCKS.
I don't want that crap swamping Linux kernel- or user-land like toxic sludge pumped up a river. FFS we already have to put up with Lennart Poe
Or just choose a different window manager (Score:2)
On Linux, it's your choice. You could run the feeble virtual window manager #2 (aka fvwm2).
Looks like Win95, last time I looked.