Or, perhaps they said “Chromium compiles easily for Linux and we can get a little goodwill from corporate IT that wants to provide managed browsing for their Linux folks”.
I don’t think the situation that the article provides is impossible, but at the same time I don’t think that Edge is a good indicator.
Yeah. Edge would be much more of an indicator if they hadn't just opted for a repackaging/repurposing of an available cross-platform engine. Nontrivial work, I'm sure, to port the Edge-specific parts of New Edge, but scarcely comparable to porting a complete legacy codebase.
Many, many enterprises still pay Microsoft billions of dollars in licenses for Windows and Office to run Excel ‘apps’ powered by VB macros. Until that has need has diminished, I really doubt Microsoft will give up their NT kernel. At least not on corporate land, maybe for consumers.
I wonder if this is why the online versions of Word, Excel, and PowerPoint remain so terrible. Online Word can't even grok some very simple types of formatting done in native Word.
I just wonder where the point of convergence will be - I imagine that even consolidation of linux and Windows will be a local minima, just like Linux as the primary Unix-ish OS (not counting apple stuff) is a local minima.
With the increasing dominance of cloud-based infrastructure, will laptops/smartphones/tablets in the 2050s all be Linux based zero-clients running on 8G networks?
>Accordingly, the return on investment of spending on Windows development is falling. As PC volume sales continue to fall off , it’s inevitably going to stop being a profit center and turn into a drag on the business.
Would that mean Microsoft could become like Apple on iDevices and force all Windows apps to be sold via the Windows Store for a 30% cut, to increase security of Windows and cut down most malware? What would be an estimate of how much money MS would have made over the years with a 30% cut similar to Apple? A couple trillion?
I think the OP is right and few people seem to realize what MS is doing by "featurizing" Linux. If devs and enterpise want FOSS and Linux (which they do) and all the enterprise users are on Windows (which they are), it makes sense for those distinctions to go away. The easiest way for MS to accomplish that is to fold Linux into the Windows desktop. Everybody gets what they want since users don't care where software runs or how.
Well at least I thought so (for the same reason as ChromeOS).
People seem divided on the topic. Obviously if distro means “Linux kernel based OS” then it’s a distro just like chromeOS. The Linux foundation for example, thinks Android is a Linux distribution.
It’s those who don’t must argue something more than just the kernel is needed, such as more GNU stuff. Exactly what that is, is what I’m curious about and why I asked.
What are the criteria required for an OS using the Linux kernel to be called a “distribution” then?
Including GNU or free software in the definition seems arbitrary...
Is there a class of applications (that aren’t apps from a store) that must be runnable?
Or do you mean that distro is short for “GNU/Linux distribution”?
I’m not arguing for/against (since it’s obviously a long-standing debate), what I’m interested in is the criteria for what makes something a distro.
Not sure if the same applies, given that BSDs don't even share the same kernel code or overall architecture, and they are created as old style UNIX, meaning each BSD caters for everything across all stack.
So you mean your definition is "distro" is slang for "GNU/Linux distribution", whereas an os using the linux kernel is not necessarily a "Linux distribution"? Or can it be a "Linux distribution" without being a "distro"?
I use Linux distributions since Slackware 2.0, it is hardly my definition, rather something that I have read somewhere on Usenet or while reading Linux Journal.
Goinng back to Android and ChromeOS.
On Android, NDK states very clerarly what are the APIs available to userspace applications written in C and C++, and none of them are Linux syscalls. Additionally everything outside of the application's installation directory is forbidden.
Starting with Android 7, Google has started to introduce security measures that kill applications that try to work around the rules.
Likewise on ChromeOS, applications are written with the Web stack, ART running on its container and same access rules, with Crostini running on top of a container, using its own kernel and talking to the ChromeOS in a way similar to WSL2 on Windows.
Hence why applications like Termux have several limitations and workarounds to fake being on Linux, and going forward they may not even work anymore, because they refuse to use Java to access the features that Google forbids direct access to the NDK.
Thanks. Requiring that userland apps can access the Linux kernel “normally” is a good definition of what it means to be a Linux OS. It seems like a much less arbitrary definition than requiring the presence of specific GNU parts.
Yeah my chromebook doesn't support it and is too weak to really run IntelliJ, I use it for web guis and some shell work. I haven't tried hard because Crostini still seems a bit bleeding edge - I think you can run linux apps directly without Crostini now too. I was waiting to see what the new PixelBook is going to do and how things settle down next year.
I suspect the reason is much more simple: it is more profitable for Microsoft to be platform agnostic. If that means supporting Linux, so be it. That does not mean they are going to abandon Windows, nor does it mean that they are going to reinvent Windows as a layer on top of Linux. It just means that they aren't picky about what their revenues are derived from.
I'm buying a new computer. Currently I have only ubuntu linux machines and was just saying to someone that I would actually consider trying windows and WSL because it would let me use the linux tools I want alongside MS office, and it might be interesting to play with windows versions of things. The reason I wont actually do it though is because I'm sure that windows will be filled with commercially oriented intrusions and annoyances that I don't see in linux. I admit I am hypersensitive to that kind of thing, but getting reminders and popups from the task bar etc destroys my concentration. Maybe there is a way to shut all that down but I always felt like windows was bothering me about stuff a lot more than linux ever does.
Definitely agree, a Mac you get Mac OS, Ubuntu you get Ubuntu, Windows you get some Frankenstein of the OS, Mcafee, Dell, and various other companies applications all stitched together.
Those parts are easily removed by doing a vanilla install the first thing you do. Windows (esp 10) still had some built-in things you don't get away from though. And also all the dialogues from self-updating software and plugins.
It is possible to disable or uninstall all the user-hostile components like Windows Update, telemetry, and the app store, but it takes some effort and manual registry editing.
The bigger problem, though, is all the legacy that's holding it back in the name of backwards compatibility. The WinAPI is older than myself. You can still run Windows 95 programs on your new Windows 10 PC. While useful to some, I'll never understand why exactly is it done this way. Apple keeps dropping old APIs to make their OSes better and simpler internally. I've never used anything that came before OS X, but I did read a lot that it was terrible and unstable, just like Windows at the time. Except modern macOS is stable and robust, while Windows still feels extremely fragile and unreliable after all these years.
Yes, I understand that Microsoft is hellbent that modern Windows versions need to be able to run 25-year-old binaries unmodified. But they could've achieved the exact same thing by putting genuine Windows 95 into a VM and seamlessly integrating it with the desktop.
> You can still run Windows 95 programs on your new Windows 10 PC. While useful to some, I'll never understand why exactly is it done this way.
Because they make a lot of money on enterprise customers. The only way to convince them to upgrade is to guarantee the new OS version won't break their legacy software. But even on consumer market, in the 90s and 2000s it was important old videogames work on new versions of Windows, or otherwise people would not upgrade.
This comes from before the days of web applications and everything as a Service. Software vendors had to work hard to get people to pay them again.
So... why upgrade in the first place then, to solve what? Something I learned as a software developer is that if a piece of code is working properly, the best thing you can do is to leave it alone.
Exactly! And that's what both the corporate executives and your parents are thinking!
So Microsoft has always had to work hard to convince people to upgrade. New features were the natural way of doing that. But not maintaining backward compatibility, in a world where almost all software was an one-time purchase, would be a deal breaker. Why upgrade if it's going to cost $20M to make our ERP suite compatible with the new Windows? Why upgrade if all the software I spent $1000 on over the past few years will no longer work? Hence backwards compatibility with existing software was a top priority (to the point that Windows has a lot of fixes to badly coded third-party software built in).
Honestly, beyond around 10.9, it's been pretty pointless from the end user perspective. Technically, they rewrote the window server to use Metal instead of OpenGL in 10.13 and that did bring performance improvements (and surfaced many bugs in the nvidia driver that they fixed in 10.14 which I'm currently on), but there were no noticeable improvements to anything since. 10.15 aka catalina just brings more restrictions on what you can run. 11.0 aka big sur is a pointless UI redesign.
Apparently, when they run out of technical things to improve and stupid code signing policies to enforce, they redesign the UI because it's a law that if you're making a popular OS it absolutely has to have a major update once a year, no matter what.
Legacy support and continued backwards compatibility is one of the biggest strengths of Windows. I can still run weird games I made 15-20 years ago without issue. As a regular users of both MacOS and Windows, I don't agree that MacOS feels more stable, recently it's felt less stable than Windows, but obviously that's anecdotal and depends on usage.
I don't see the problem with GP's suggestion. Couldn't you still run your 15-20 year old games if Windows 10 shipped with a Windows XP VM that integrated seamlessly with the Desktop?
It's possible that performance would be an issue with some use cases, startup times, etc. Eventually this will probably be a reality, in the same way you can't run DOS binaries without an emulator or VM
i booted into my windows machine the other day, and when it finally came up, it was full of crap: my start bar had a bing ad for a halloween quiz, 3 different versions of candy crush, half my settings reset, etc... compared to my daily (linux and macos) it sure felt like i had little control over the machine, it was a microsoft ad platform first and place to get actual work done second
I use Windows 10 with Classic Start Menu. I see no adverts. I do see some notifications, almost all of which are useful. Probably you should try things first, before jumping to conclusions - I find Windows 10 a pretty good experience, and yes, I have also been using Linux for about 30 years.
So do I .. however recently Windows has been causing Classic Start Menu upgrade commands to fail - Classic still seems to work, but I don't know for how long.
I think this is a trend of MS breaking the tools we use to get some control back over Windows, because they've broken some of my others recently also, such as Windows defender flagging Spybot S&D's immunization as a hosts hijack. And of course there's no easy way to work around that, pressing ignore didn't work it produce the same result as quarantine .. I'm so angry at MS for all the shit they push, but there's just not a better alternative they closely walk the line of costs/benefits .. which is why I fear that as their position on the desktop strengthens with things like wide WSL adoption, we will see more of their anti-features and domineering breakages of the tools people write to resist them.
- don't use online account, but offline (local) one
- use WPD to uncheck absolutely everything related to windows telemetry / sending info. I keep only updates and those set to manual.
- and can't stress this enough - get a PiHole, either on RPi or re-purpose a crappy old laptop. This one is also good for your smartcrap you usually get to your spouse/kids
And the "Created and mainteined by" links to maintainers' Steam profile... I'm sorry and I'm sure this is my biased view but gamers usually make very questionable choice on security stuff even they have 0 malicious intent.
In addition to that all that, I once caught Windows Defender submitting places.sqlite out of my Firefox profile directory so I'd recommend you turn off "automatic sample submission" in Windows Defender if WPD doesn't already do that.
As a Linux desktop user, I've only ever been able to tolerate LTSC releases of Windows 10, and those still have telemetry features. They also lag behind Windows 10 releases so that things like WSL2 are still unsupported.
I tried running Windows as my daily driver because I was somewhat sold on WSL and WSL 2, however using non-LTSC releases of Windows is a frustrating experience, and there were too many incompatibilities between enterprise Windows and the desktop/development environment I was trying to use it as.
My solution is to have an LTSC virtual machine set up so that I can dip into Windows when I need to, and just use a sane Linux environment as the host OS.
it would let me use the linux tools I want alongside MS office
I used to go this route. I used WABI (Windows Application Binary Interface) on Solaris in the 1990s, sort of like a very early WINE. However as time has gone on, I found that the Linux Office Suites (LibreOffice, more recently) have become more and more sufficient for my needs.
I realised recently that I haven't really booted Windows itself more than half a dozen times over the last few years.
I got my first Windows 10 a month or two ago with a new laptop. I then proceeded to squeeze it down to under 100 gig on a 1 terabyte hard drive. On my main desktop, Windows 8 is relegated to a mere 50 gig out of 12 terabytes of storage.
I have used Windows 7 on VirtualBox more often for running my HP Scanner's MultiFeed software, just because I can. Though normally, I just run a single-sheet scan most of the time under Linux.
In summary, Windows could completely disappear and it would not affect me adversely at all.
I tried to use WSL to run a cron job, but that wasn't supported as of 2019. I wasn't able to figure out how to schedule and run arbitrary Python code on Windows.
But Windows lets you do things no Linux distro really can - use modern, state-of-the-art interface technologies like multitouch, pens, and dials. A Surface with WSL is about the ideal setup, with the best of both worlds.
Also, the author of this article seems ignorant of the fact that Windows' kernel internals are one of its strongest features - Cutler's NT Kernel incorporated many things he learned from building VMS, and while the shell and outer parts of Windows have often left a lot to desire, the NT (not DOS/Win9X) kernel is and always has been excellent. I don't see Windows evolving to run on a Linux kernel - that would be only viable with vast improvements/changes to the Linux kernel.
> Also, the author of this article seems ignorant of the fact that Windows' kernel internals are one of its strongest features ...
I would not be surprised at all if he's ignorant of Windows' kernel internals or, really, much else about Windows.
> [...] ESR, is an American software developer, open-source software advocate, and author of the 1997 essay and 1999 book The Cathedral and the Bazaar. -- https://en.wikipedia.org/wiki/Eric_S._Raymond
In fact, this article likely would have gained little attention were it written by anyone else.
This argument totally ignores Microsoft’s consideration of what might happen after that endgame-state, and what utility (if any) it would have for them. Specifically, Microsoft would have everything to lose if it placed all of the Windows-like APIs into their Linux kernel, even if that were in the interests of “de-complexifying” their Azure business, because after that they’d be on par with every other Cloud/PaaS provider and they’d be outcompeted on cost by the likes of Amazon Web Services.
I think this would have been obvious when Microsoft pushed a "free" upgrade to every Windows 7 Home PC, to bring it to Windows 10. What they care about now are recurring revenues from monthly subscription services (xbox live gold type things), office365 type things, and azure services. And all of the usual tricks used to get people to use Edge as their default browser. Giving away the OS for free to accomplish that is clearly the way they're going about that.
I am afraid of EEE strategy of Microsoft . But,Windows having WSL means that Developers can target Linux only without even considering Windows , being sure that all Linux binaries run on WSL . But proprietary extensions of WSL should never ever be used.
EEE is not going to have any impact on Linux (and also, EEE hasnt been Microsof's strategy for a while). Once they start extending with proprietary additions, you basically just end up where you were with two operating systems, a proprietary windows and an open source Linux. They can't force the kernel to accept proprietary extensions.
> They can't force the kernel to accept proprietary extensions.
They don't need to. They just need to offer a "better" Linux (kernel and/or distro) that also runs Excel, and is backed by a global Fortune 10 corporation.
It's Linux, it's legal, it's E₂ and brings E₃ within reach if they choose that path.
There is a concern that mainstream proprietarized "Linux" becomes completely incompatible with real Linux to the detriment of Linux's values and the people who care about those values.
>But,Windows having WSL means that Developers can target Linux only without even considering Windows
I've been using WSL1 and WSL2 regularly since they were in beta. They have both exceeded my expectations. I still SSH into my local development box I use for webdev and the like because it's still faster, less limited and more powerful than WSL/2. Again, I think WSL is awesome and I use it all the time - to do a few linuxy things here and there on my windows box. It's nice, but it is in no way a production-grade replacement for native linux.
> "Ten years later, Azure makes Microsoft most of its money."
Well, no, not yet. Microsoft doesn't break out Azure separately. It's usually included in their "Intelligent Cloud", which is the largest segment, but that also includes SQL Server (a big money maker), Windows Server, GitHub, and other stuff. "Commercial Cloud" revenue (which is all the cloud portions of their products, from Azure to Office 365 Commercial) was $51.7Bln of $143Bln overall. So yes, Azure makes a lot of money, but not "most".
Microsoft is talking about swapping out the windows portion of their azure stack for Linux. Still running their own hypervisor, but with Linux as the "dom0" controlling it.
> Microsoft doesn't break out Azure separately ... "Commercial Cloud" revenue (which is all the cloud portions of their products, from Azure to Office 365 Commercial)
I'm begging people, please stop repeating this. Microsoft has been reporting O365 revenue separately from Azure revenue for like five years now.
Whichever way you spin it, Microsoft isn’t making much money from desktop Windows licenses.
And if they did move to Linux you’d still need CALs for all the proprietary stuff on top, so it’s unlikely they’d be losing much, if any money by doing this.
I agree that we'll see the boundaries between Windows and Linux blur, but I don't think we'll ever see Microsoft give up the decades of backwards compatibility that have kept people on Windows since the beginning. There will always be demand for running older Windows apps (games, legacy business software, etc..) on a modern OS, and I don't think Microsoft will cede that space to a competitor (and with good reason!) It will likely continue to shrink as a proportion of revenue, but it will be worth much more than that to keep it as a toehold to deliver other software and services: Xbox game pass, visual studio, azure services, etc...
What about Linux distro virtual machines? It seems as though some users prefer running a virtual machine on different hardware rigs. I'm not sure if this will increase in popularity, yet it appears as a manageable solution.
I don't think MS is porting Edge to Linux because they want to test the waters for running Windows on a Linux kernel as the author speculates. It seems much more likely that Microsoft recognizes the web is an important platform for them to be a player on in the future, and they want web developers to be able to test stuff on Edge regardless of which OS they use.
This is the reason why Edge is available on MacOS and Linux.
I saw a previous comment mention Chrome OS as improving in terms of market share. Honestly, running a Debian virtual machine over Chrome OS appears to remain a lightweight desktop environment, especially without extensive hardware configurations.
I did the opposite. After switching to OSX 16 years ago, I am switching to Windows.
Either way I have been a Linux user since 1992.
Reasons:
* WSL1/2
* Better, cheaper laptops and desktops with better keyboards, F-keys, more GPUs
* No wondering when the next thing Apple will do to remove something I use (F-keys, ESC, 32-bit, ability to actually do stuff on my own filesystem, etc.) or break something that just worked (like Homebrew)
...
This is the Android problem all over again. Some argue that Linux is the most popular graphical desktop/mobile OS. I think, however, that back then when we were dreaming about Linux desktops we wanted the freedom, not necessarily or only the kernel itself.
So if Linux and Unix end up dominating everything in the form of Android, MacOS and Windows, that's not a huge win for the FLOSS community.
> Microsoft has already ported Edge to run under Linux. There is only one way that makes any sense, and that is as a trial run for freeing the rest of the Windows utility suite from depending on any emulation layer.
This is a very strange conclusion to draw. Chromium was already ported to Linux. Microsoft just "ported" the Edge skin which likely uses very little Windows-specific code.
> and Linux finally wins the desktop wars, not by displacing Windows but by co-opting it
That's not a "win", though. That's "being subsumed". After all, if LSW works, why would anyone with normal computing needs ever need to install Linux on a computer ever again? Or even in a VM, for that matter? As long as they have Windows, they get the parts of Linux they need for free, without the headache of actually learning the convoluted finer points of properly using a well configured Linux distro.
If anything, Linux finally loses the desktop war, by no longer being needed as a standalone OS. It's now something you just fire up on the side if you need something that only has a Makefile, rather than an .sln file...
Your desktop is still Windows, your applications are still Windows, and LSW is there for when you need to do "that thing that only people who work with computers for a living need to do". No one even needs to be told that when they run "bash", they're running Linux: it's now just another app you run. Linux didn't just lose, it got made irrelevant.
And that's another thing the vast majority of folks this is for couldn't care less about. It's all the same "make it irrelevant, this is not something I should need to care about to do whatever I need WSL for".
I don't necessarily see it as a "win" for Linux, but I do see it as an admission of defeat for windows.
If windows (and it's related tooling) is so garbage that you have to literally bundle an entire OS and tools in it just so that a developer can get real work done, then that speaks volumes.
And also why would I want all of the pain of using windows, (with linux bundled in it) if I can just use linux?
Never needed linux until xyz tool chain that is not crossplatform require linux to compile...
I welcome WSL only because it will be easier to test my apps against linux.
164 comments
[ 2.7 ms ] story [ 230 ms ] threadI don’t think the situation that the article provides is impossible, but at the same time I don’t think that Edge is a good indicator.
With the increasing dominance of cloud-based infrastructure, will laptops/smartphones/tablets in the 2050s all be Linux based zero-clients running on 8G networks?
Would that mean Microsoft could become like Apple on iDevices and force all Windows apps to be sold via the Windows Store for a 30% cut, to increase security of Windows and cut down most malware? What would be an estimate of how much money MS would have made over the years with a 30% cut similar to Apple? A couple trillion?
That is except if the US court messes up in the ethic vs. Apple case giving Microsoft a implicit green lighting for such a thing.
https://en.wikipedia.org/wiki/Eric_S._Raymond#Political_beli...
ChromeOS is storming along - best desktop version of linux around. If I could easily run IntellJ on it I would switch.
It's not a Distro.
I thought any packaging of Linux kernel + whatever surrounding software needed to make an OS is a “distro”?
People seem divided on the topic. Obviously if distro means “Linux kernel based OS” then it’s a distro just like chromeOS. The Linux foundation for example, thinks Android is a Linux distribution.
It’s those who don’t must argue something more than just the kernel is needed, such as more GNU stuff. Exactly what that is, is what I’m curious about and why I asked.
Android and ChromeOS are Marketplace-(App) Runtime NOT a Linux Distro
>The Linux foundation for example, thinks Android is a Linux distribution
Well they compare it with embedded Linux:
https://www.linuxfoundation.org/blog/2013/03/intro-to-embedd...
> Wikipedia and others list it as a distribution.
No they don't:
>>Android is a mobile operating system based on a modified version of the Linux kernel and other open source software
Debian:
>>Debian (/ˈdɛbiən/),[5][6] also known as Debian GNU/Linux, is a Linux distribution composed of free and open-source software,
https://en.wikipedia.org/wiki/Linux_distribution#Linux_kerne...
https://en.wikipedia.org/wiki/Linux_distribution
Including GNU or free software in the definition seems arbitrary... Is there a class of applications (that aren’t apps from a store) that must be runnable?
Or do you mean that distro is short for “GNU/Linux distribution”?
I’m not arguing for/against (since it’s obviously a long-standing debate), what I’m interested in is the criteria for what makes something a distro.
Neither ChromeOS nor Android expose the Linux kernel directly to userspace apps, so for them is irrelevant what kernel is being used.
Or BSD ;)
Why is that not a Distribution, it's not a Gnu/Linux Distribution true, but it's one for sure. You know what BSD means?
Berkeley Software Distribution
>meaning each BSD caters for everything across all stack
Like every "real" Linux Distribution
So you mean your definition is "distro" is slang for "GNU/Linux distribution", whereas an os using the linux kernel is not necessarily a "Linux distribution"? Or can it be a "Linux distribution" without being a "distro"?
Goinng back to Android and ChromeOS.
On Android, NDK states very clerarly what are the APIs available to userspace applications written in C and C++, and none of them are Linux syscalls. Additionally everything outside of the application's installation directory is forbidden.
Starting with Android 7, Google has started to introduce security measures that kill applications that try to work around the rules.
Likewise on ChromeOS, applications are written with the Web stack, ART running on its container and same access rules, with Crostini running on top of a container, using its own kernel and talking to the ChromeOS in a way similar to WSL2 on Windows.
Hence why applications like Termux have several limitations and workarounds to fake being on Linux, and going forward they may not even work anymore, because they refuse to use Java to access the features that Google forbids direct access to the NDK.
So hardly a Linux distribuition.
Sure it is.
It's not a general purpose distro (running on GP/COTS hardware), but it is a software distribution built around a Linux kernel.
No google Play Store is the distributor NOT Android, Android and Linux is just part of the runtime.
If you don't mind surrendering yourself to Google, at least.
Solve that and I may be a convert.
And how!
The bigger problem, though, is all the legacy that's holding it back in the name of backwards compatibility. The WinAPI is older than myself. You can still run Windows 95 programs on your new Windows 10 PC. While useful to some, I'll never understand why exactly is it done this way. Apple keeps dropping old APIs to make their OSes better and simpler internally. I've never used anything that came before OS X, but I did read a lot that it was terrible and unstable, just like Windows at the time. Except modern macOS is stable and robust, while Windows still feels extremely fragile and unreliable after all these years.
Yes, I understand that Microsoft is hellbent that modern Windows versions need to be able to run 25-year-old binaries unmodified. But they could've achieved the exact same thing by putting genuine Windows 95 into a VM and seamlessly integrating it with the desktop.
Because they make a lot of money on enterprise customers. The only way to convince them to upgrade is to guarantee the new OS version won't break their legacy software. But even on consumer market, in the 90s and 2000s it was important old videogames work on new versions of Windows, or otherwise people would not upgrade.
This comes from before the days of web applications and everything as a Service. Software vendors had to work hard to get people to pay them again.
So Microsoft has always had to work hard to convince people to upgrade. New features were the natural way of doing that. But not maintaining backward compatibility, in a world where almost all software was an one-time purchase, would be a deal breaker. Why upgrade if it's going to cost $20M to make our ERP suite compatible with the new Windows? Why upgrade if all the software I spent $1000 on over the past few years will no longer work? Hence backwards compatibility with existing software was a top priority (to the point that Windows has a lot of fixes to badly coded third-party software built in).
Hand on heart, do you really feel OS X/macOS is improving much?
Apparently, when they run out of technical things to improve and stupid code signing policies to enforce, they redesign the UI because it's a law that if you're making a popular OS it absolutely has to have a major update once a year, no matter what.
I think this is a trend of MS breaking the tools we use to get some control back over Windows, because they've broken some of my others recently also, such as Windows defender flagging Spybot S&D's immunization as a hosts hijack. And of course there's no easy way to work around that, pressing ignore didn't work it produce the same result as quarantine .. I'm so angry at MS for all the shit they push, but there's just not a better alternative they closely walk the line of costs/benefits .. which is why I fear that as their position on the desktop strengthens with things like wide WSL adoption, we will see more of their anti-features and domineering breakages of the tools people write to resist them.
I don't need to smoke crack to know that I don't want anything to do with it.
- don't use online account, but offline (local) one
- use WPD to uncheck absolutely everything related to windows telemetry / sending info. I keep only updates and those set to manual.
- and can't stress this enough - get a PiHole, either on RPi or re-purpose a crappy old laptop. This one is also good for your smartcrap you usually get to your spouse/kids
https://wpd.app/
As a Linux desktop user, I've only ever been able to tolerate LTSC releases of Windows 10, and those still have telemetry features. They also lag behind Windows 10 releases so that things like WSL2 are still unsupported.
I tried running Windows as my daily driver because I was somewhat sold on WSL and WSL 2, however using non-LTSC releases of Windows is a frustrating experience, and there were too many incompatibilities between enterprise Windows and the desktop/development environment I was trying to use it as.
My solution is to have an LTSC virtual machine set up so that I can dip into Windows when I need to, and just use a sane Linux environment as the host OS.
I used to go this route. I used WABI (Windows Application Binary Interface) on Solaris in the 1990s, sort of like a very early WINE. However as time has gone on, I found that the Linux Office Suites (LibreOffice, more recently) have become more and more sufficient for my needs.
I realised recently that I haven't really booted Windows itself more than half a dozen times over the last few years. I got my first Windows 10 a month or two ago with a new laptop. I then proceeded to squeeze it down to under 100 gig on a 1 terabyte hard drive. On my main desktop, Windows 8 is relegated to a mere 50 gig out of 12 terabytes of storage.
I have used Windows 7 on VirtualBox more often for running my HP Scanner's MultiFeed software, just because I can. Though normally, I just run a single-sheet scan most of the time under Linux.
In summary, Windows could completely disappear and it would not affect me adversely at all.
Also, the author of this article seems ignorant of the fact that Windows' kernel internals are one of its strongest features - Cutler's NT Kernel incorporated many things he learned from building VMS, and while the shell and outer parts of Windows have often left a lot to desire, the NT (not DOS/Win9X) kernel is and always has been excellent. I don't see Windows evolving to run on a Linux kernel - that would be only viable with vast improvements/changes to the Linux kernel.
I would not be surprised at all if he's ignorant of Windows' kernel internals or, really, much else about Windows.
> [...] ESR, is an American software developer, open-source software advocate, and author of the 1997 essay and 1999 book The Cathedral and the Bazaar. -- https://en.wikipedia.org/wiki/Eric_S._Raymond
In fact, this article likely would have gained little attention were it written by anyone else.
They don't need to. They just need to offer a "better" Linux (kernel and/or distro) that also runs Excel, and is backed by a global Fortune 10 corporation.
It's Linux, it's legal, it's E₂ and brings E₃ within reach if they choose that path.
I've been using WSL1 and WSL2 regularly since they were in beta. They have both exceeded my expectations. I still SSH into my local development box I use for webdev and the like because it's still faster, less limited and more powerful than WSL/2. Again, I think WSL is awesome and I use it all the time - to do a few linuxy things here and there on my windows box. It's nice, but it is in no way a production-grade replacement for native linux.
Well, no, not yet. Microsoft doesn't break out Azure separately. It's usually included in their "Intelligent Cloud", which is the largest segment, but that also includes SQL Server (a big money maker), Windows Server, GitHub, and other stuff. "Commercial Cloud" revenue (which is all the cloud portions of their products, from Azure to Office 365 Commercial) was $51.7Bln of $143Bln overall. So yes, Azure makes a lot of money, but not "most".
I would assume Microsoft sees competitive advantage in using their own OS and hypervisor.
Office is part of “ Productivity and Business Processes” not “Intelligent Cloud”.
https://www.cnbc.com/2019/10/23/microsoft-msft-earnings-q1-2...
I'm begging people, please stop repeating this. Microsoft has been reporting O365 revenue separately from Azure revenue for like five years now.
And if they did move to Linux you’d still need CALs for all the proprietary stuff on top, so it’s unlikely they’d be losing much, if any money by doing this.
This is the reason why Edge is available on MacOS and Linux.
And, notably, Linux is _not_ Unix, for a variety of reasonable reasons.
Either way I have been a Linux user since 1992.
Reasons:
* WSL1/2 * Better, cheaper laptops and desktops with better keyboards, F-keys, more GPUs * No wondering when the next thing Apple will do to remove something I use (F-keys, ESC, 32-bit, ability to actually do stuff on my own filesystem, etc.) or break something that just worked (like Homebrew) ...
So if Linux and Unix end up dominating everything in the form of Android, MacOS and Windows, that's not a huge win for the FLOSS community.
I’m not sure how anyone could arrive at this conclusion. The OS itself is certainly not “Linux underneath a carefully preserved old-Windows UI.”
This is a very strange conclusion to draw. Chromium was already ported to Linux. Microsoft just "ported" the Edge skin which likely uses very little Windows-specific code.
That's not a "win", though. That's "being subsumed". After all, if LSW works, why would anyone with normal computing needs ever need to install Linux on a computer ever again? Or even in a VM, for that matter? As long as they have Windows, they get the parts of Linux they need for free, without the headache of actually learning the convoluted finer points of properly using a well configured Linux distro.
If anything, Linux finally loses the desktop war, by no longer being needed as a standalone OS. It's now something you just fire up on the side if you need something that only has a Makefile, rather than an .sln file...
Your desktop is still Windows, your applications are still Windows, and LSW is there for when you need to do "that thing that only people who work with computers for a living need to do". No one even needs to be told that when they run "bash", they're running Linux: it's now just another app you run. Linux didn't just lose, it got made irrelevant.
As for your bash example, I have run various versions of it across many kinds of POSIX like OSes.
Yes Linux is irrelevant, it is just a free beer UNIX clone, tomorrow another UNIX clone can take its place.
WSL is about fixing that little mistake, while taping on Linux ABI as extra.
If windows (and it's related tooling) is so garbage that you have to literally bundle an entire OS and tools in it just so that a developer can get real work done, then that speaks volumes.
And also why would I want all of the pain of using windows, (with linux bundled in it) if I can just use linux?