Hacker News new | ask | show | jobs
by esics6A 1678 days ago
This is true and Microsoft is known to be a company that has many product groups that are often in direct conflict with each other.

Cloud Services is the "cool" Microsoft and their cloud services are now the biggest profit center for them. They have Azure, developer tools and contribute a huge amount to the overall open-source world. The majority of their customers run Linux workloads and things like Kubernetes. Cloud folks at Microsoft don't care if you run Linux or MacOS as long as you run your web apps on Azure that's all they care about and they try to make it as easy as possible. These are the guys that developed Microsoft's first internal Linux distribution.

Satya Nadella comes from the Cloud/Azure part of the business. That's why Microsoft has been doing some really innovative and smart choices lately under his leadership.

The Windows group are the luddites at Microsoft. Instead of open-sourcing Windows and treating it as open technology something the Cloud group would love they still jealously guard it as proprietary. They only put WSL (Windows Subystem Linux) because the Cloud group wanted it. Windows group is living in the 1990's and still thinks you can make money on a desktop operating system. They put ads and use dark patterns because they're desperate to show that Windows can still make money. Their CEO is now one of the Cloud guys and for them that sucks.

10 comments

> Windows group is living in the 1990's and still thinks you can make money on a desktop operating system.

I can't speak for others but I'm more than willing to pay good money for a version of Windows that isn't actively user hostile and full of garbage. Like, give me a Windows 2000 UI on top of modern Windows underpinnings and I'll give you $500/year. They can even keep putting out the shitty version of Windows and just make it explicitly ad/creepy-data-harvesting supported. That might still not be enough to actually make them money, I can't say.

What I can say is that I feel like Window's current strategy is less about making money and more about actively trying to kill off personal computing for the sake of pushing everyone to subscription services.

I typically get downvoted for mentioning this, no idea why: This is basically Windows Enterprise LTSC/LTSB.

That is what I use and that it what I set up for my less technically competent friends and family who are not good at navigating the ever-shifting sands of the Windows UI, the situation changing on them, a game they had not remembered installing appearing in their Programs. Working on the simple rhythm of muscle memory and repetition, they draw the Worm.

> I typically get downvoted for mentioning this, no idea why: This is basically Windows Enterprise LTSC/LTSB.

In the context of a discussion about Microsoft’s intentions, it is very clearly asshole behavior by Microsoft to have a perfectly good product that people want, but then put up artificial barriers like having a friend who knows how to get it and install it.

I would not downvote you for mentioning it, but it makes Microsoft look even worse than if they did not have LTSC at all since it removes their plausible deniability.

I was able to use LTSB/LTSC legally for years: forget it. A lot of stuff won't run anymore. Even third-party software (e.g. nextcloud client).
I am a power windows user and its such a shame ( on my end ) that I did not have a clue that Windows 10 LTSC existed till recently. I was quite happy with the Windows 10 PRO version, wanted to check out WIN 11, installed it, it started giving me the dreaded memory leak issue ( on AMD currently ) when opening windows explorer. Tried to revert back to windows 10 but the option was greyed out ( 10 days is the max time to revert back to Windows 10 ). Got to know about Windows 10 LTSC version, activated it with KMS and its snappy as f, no Ads, no Microsoft Store, no cortana, no bullshit. Basically the version people should've gotten in the first place.
> 10 days is the max time to revert back to Windows 10

What, why?!

Yes, after upgrading to windows 11, you get a 10 day period during which you can revert back to windows 10. After that it greys out automatically. I was forced to do a fresh install because of how laggy windows 11 was.
What's the purpose of this?

Means it's TECHNICALLY possible, and Microsoft actively takes away the option after 10 days?

Do people not hate this?

On macos the only way to revert is through full machine backups. That at least lets me think the upgrade is irreversible without backups, and then I am fine with it. :)

While I love LTSC, I think it is not possible to use the windows subsystem for linux on it.. which is my only pain point.
Sure is, on a new 2021 one.
Obtaining Windows Enterprise legally is tricky if you're not a big corporation. It's not even a money thing, it's that they outright won't sell you a license directly, instead you have to go to a reseller and then pad the order with cheap client access licenses to reach the minimum amount of licenses required for an order.
Here is a relevant thread that goes into how complicated it is to purchase:

https://community.spiceworks.com/topic/2167558-explicit-inst...

What if you subscribe to technet (or whatever they call it now ? MSDN ?) ... don't you get one license to everything ?

The last time I paid for MSDN was 2011 or 2012 but I was able to download and use everything ...

Visual Studio (formerly MSDN) subscriptions come with purpose-specific (for software development and/or testing, depending on the subscription) licenses for lots of things (exactly what also depends on the tier, IIRC).

The usage constraints are mostly social/legal rather than technical, of course.

Visual Studio subscription is effectively $200/yr if you use Azure at all. It's $800/yr, but you get a $50/month Azure credit. It's not hard to use $50/month worth of Azure stuff.
You can license the enterprise upgrade if you have an o365 subscription.
The problem with LTSC is that newer games won't work, other than that it's a pretty reasonable experience. I put vanilla Win 10 on computer recently and it's bad. Luckily the games I'm trying to run are better supported on linux via proton than they are on LTSC, and there's no way in hell I'm installing win 11, I already find the lack of proper programmatic file-association support infuriating.

I think at least personally I've finally reached the tipping point where dealing with the annoyance of running windows apps on linux is going to be worth not having the huge host of annoyances microsoft is intentionally pushing onto me.

LTSC still contains a hefty chunk of bullshit, in my opinion, including all the god awful unnecessary UI changes they only ever half implemented and the laggy apps that take ages to launch and use up entirely too much memory.
Oh, it isn't perfect by any means. I still have to do a lot of tweaks, but the whole forced update thing ... makes me insane.
> I typically get downvoted for mentioning this, no idea why: This is basically Windows Enterprise LTSC/LTSB.

Because it's incredibly hard to outright impossible to get a legal version of this. AFAIK it's only available for MSDN subscribers, OEMs/ODMs/embedded device manufacturers and large-volume customers and illegal to resell; on top of that unlike regular Windows ISOs there are no public download links that you can then either activate with a key you happened to find somewhere or run one of the usual activation cracks.

No matter what, unless you are in a highly privileged position you do not have a way of obtaining LTSB without exposing yourself to legal or security risks.

ETA: Just had a look on a well-known torrent site - "LTSB" yielded no usable results (> 10 seeds), and "LTSC" only one that matches the seed count, and it doesn't even ship supposedly virgin ISOs but modified ones, so no way to know if at least the install media is free of malware (by comparing it to known MS hashes). Jesus, I didn't know the situation was that bad.

That's weird, because I have used Bittorrent once, about a decade back, just to see what it was like, for a few movies but no software because, well, malware. I downloaded ISOs from a MS site and bought a key off of a reseller. So, no, I don't feel highly privileged.
> I downloaded ISOs from a MS site and bought a key off of a reseller.

The latter part is the problem. LTSB/C keys are illegal to resell and MS regularly bans keys they suspect of having been sold.

Can you install and play video games on LTSC? Is there any multimedia limitation in that regard?
The fact that they are killing Office just so that they try to push people into the cloud should be enough evidence that they have this strategy.

Compared to MS Office, Windows is nothing.

> they are killing Office

There is no such plan.

The paid pro version still has the ads and upselling tricks.

Its almost comical.

Aren't you asking for Windows Server?
No. Windows server w/ Desktop Experience contains pretty much all the same bullshit as Windows 10. Even LTSC only cuts out a relatively small portion of it.
Sounds like Windows Server... 2008
Maybe I'm overly paranoid (but considering Microsoft's behavior over the past twenty years, it would be in character), I can't help but think that WSL is a stalking horse for forbidding Linux from bare metal booting on consumer PCs. Revoke the Linux distro UEFI keys for "reasons" and there you have it.
That's not paranoid at all. MS goal is to declare Linux a "legacy" environment, like they tried to do with UNIX in the 80s (the so called POSIX subsystem).
The part that really pains on this, is it is a common pattern in our entire industry. :(
Secure boot could be a good feature, but it can be severely abused. Some people really put effort in it to make it more secure. But the usual problems of certificate logistics leads to the case the Microsoft basically owns secure boot.

Only a matter of time before Netflix demands remote attestation.

>Only a matter of time before Netflix demands remote attestation.

Which is completely absurd. Within minutes of Netflix/anyone releasing anything new I can download it via Usenet or torrents. Nothing anyone does will change that reality. They just continue to make it more difficult for the average user to access content via player requirements for various DRM schemes. Which funny enough pushes people back to pirating to avoid all the nonsense.

It must be soul crushing to work on DRM for any industry as your job. Knowing that no matter what you do, how clever you are, etc - it doesn't matter at all, it'll be bypassed very shortly.

> It must be soul crushing to work on DRM for any industry as your job. Knowing that no matter what you do, how clever you are, etc - it doesn't matter at all, it'll be bypassed very shortly.

I can't imagine doing this, to be honest if I were asked to implement DRM I would quit, or try to make sure there are fundamental flaws in the scheme, it's the only moral thing to do if you're put in that situation imo.

EDIT: To those upvoting me, software as a service is usually DRM, so maybe hold your upvote if you're complicit.

> Only a matter of time before Netflix demands remote attestation.

They already do on Android. Root your device and either play cat and mouse with Magisk and Universal SafetyNet fix, or get downgraded to L3. And even then, it's uncertain how long the USN hack will last, given that all it does is pretend the device doesn't have hardware attestation - once Google decides to mandate HA for all Android 12 and above devices (as all SoCs capable of running that should have some form of secure element), it's game over.

We need legislative action against anti-rooting measures and other shit that takes away control of devices from users, and that fast, but it doesn't look like it's high on the priority list for the next year of the Biden admin, and after that we will likely see, once again, a gridlocked Congress and in 2024 the Rise of the Sith again.

> Only a matter of time before Netflix demands remote attestation.

Well, as soon as they do that they become useless for me and I'll unsubscribe.

I don't know how many people are on the same boat, but between badly developed smart tvs and top boxes there are probably a large number.

Bye, Netflix.
I think of Windows as a legacy environment, having abandoned it when Windows 7 ended. That's when Microsoft exited the operating system business and entered the ad business.
> MS goal is to declare Linux a "legacy" environment

Your information is a decade out of date. More than half of Azure instances run Linux, and it's a huge source of income for MS now. Billions huge.

Microsoft is a very large company. Anyone who's worked at a company even a tenth the size of Microsoft knows how hard it is to get the company to pick a direction.

The only real direction you can get everyone to agree on is "profit," and the nice thing about the existence of Azure as one of their major profit centers is that Microsoft now has a bunch of business that depends on Linux running well. A company mostly running Linux is already looking at them as third place behind Amazon and Google - if Microsoft risks making kernel developers unable to run Linux, they risk making kernel developers unwilling to accept patches to make Azure (or WSL) run Linux well. I'm not saying they're never going to try it, since the Azure org doesn't control the UEFI signing program, but I am saying there's a significant part of the company that will say that it threatens their profits to a scale much larger than the lost Windows licenses on consumer PCs.

Fun fact: every single example in the Wikipedia "embrace, extend, and extinguish" article https://en.wikipedia.org/wiki/Embrace,_extend,_and_extinguis... is something that failed (IE is dead, Office 365 works great on non-MS browsers, MS has no influence on Java, MSN Messenger is dead, MAPI is dead, etc.). The only thing harder than getting a company the size of Microsoft to agree on something is to get the rest of the industry to go along with it too. The few cases where they've succeeded (e.g., the thing from the 1980s where they made OEMs pay them even for machines that didn't ship with MS-DOS) were when they had a sufficiently technically superior product they could use to bully people. They don't have that power anymore.

> Fun fact: every single example in the Wikipedia "embrace, extend, and extinguish" article https://en.wikipedia.org/wiki/Embrace,_extend,_and_extinguis... is something that failed (IE is dead, Office 365 works great on non-MS browsers, MS has no influence on Java, MSN Messenger is dead, MAPI is dead, etc.).

You're assuming the goal of these things was to replace the product, when the goal was to destroy the competitor.

Netscape is dead (and Firefox is dying), all the historical competitors to MS Office are dead or have negligible market share, Java never got enough market share to make it easy for people to switch away from Windows, AIM/ICQ/whatever are all dead (and worse, nobody really uses XMPP or other open protocols), hardly anybody runs their own email server anymore, etc.

They successfully destroyed all of the independent versions of these things, so that now their competitors are only the likes of Google and Facebook who get where they are by leveraging their own dominant market positions in other markets.

It's also kind of disingenuous to say "IE is dead" and "MSN Messenger is dead" and ignore Edge and Skype and Teams etc.

Windows market share is the lowest it's ever been. Haven't used windows myself since March 2020 and nothing is making me want to change that.
As far as I am concerned, the whole future of Microsoft is Azure and extracting rent from Office.
Well, until the LibreOffice Foundation gets a bit smarter or greater business support from some heavy hitters. Microsoft relying on Office is on shaky ground.

People arguably use MS Office out of habit and since it's a forced de facto standard at work. But, there is no reason they can't switch to LibreOffice for 95% plus of what people are doing.

And they can't even get those right, between Skype and Teams it would be hard to pick the bigger failure.
Underestimating the JVM there a bit. Sucks that Oracle tried to monster it, but it's a solid workhorse. Also, down with MS, leopards don't change their stripes so easily, a change of helm won't turn a company with greed so deeply entrenched in their DNA. *Spots not stripes. Or tigers.
So this requires some context.

Java was released by Sun Microsystems, a Unix vendor with their own hardware architecture. The big feature was "write once, run anywhere." Write the program once and it would run on Windows/Intel as well as Solaris/SPARC.

The Java language wasn't horrible and WORA was a big advantage in the days when there were half a dozen Unix vendors plus Mac and Windows and Novell Netware etc. So it was becoming popular, and every app the developer decided to write in Java was one that wasn't tied to the Win32 API and therefore Windows.

It's also possible to compile other languages to Java bytecode as along as they didn't use OS-specific APIs -- and Java provided platform-independent ones. So there was a real risk that everything would end up running on a platform-independent JVM. Which Microsoft successfully prevented from happening for long enough for Sun to run out of money and get consumed by Oracle.

It is hard to remember, or for younger developers to believe, but there was a time in the mid 90s when:

1. Most programing languages sucked

2. Java was fresh and new

Java had a lot going for it. It was free, with a functioning IDE that had a working graphical debugger! MS had just about finished killing Delphi (which also cost $$) and over in *nix land the GUI libraries were fighting amongst themselves and Linux wasn't something even an average developer was going to install.

So you had Perl, raw C, the horrors that were DSLs and frameworks written in the c++ of the time, then Java came along.

Applets failed, sure. And back then everyone wanted their GUIs to look like the platform native UI (how times have changed!) but Swing was super easy to write UIs in.

I'd wager the majority of programmers, outside of ex-LISP folks, didn't know what closures were, and functions as first class objects wasn't on anyone's mind.

So all of Java's shortcomings didn't seem like a big deal. It "compiled" fast, had actual packages you could distribute and import easily, and the compiler errors made sense.

So yeah Java was a fresh breath when it came out.

Then c# came out a bit later and was basically better in a million small ways from day 1, except it wasn't open source so a community never built up around it in the same way.

Now days we are spoiled for languages to choose from.

None of those examples have anything to do with "embrace, extend, extinguish"? And they're misleading at best.

Netscape died because browsers became a part of the baseline operating system. That wasn't Microsoft's choice, that's just de-facto reality (see: every other OS). Just imagine shipping a computing device today without a browser! Steve Jobs put it best (to a different company): You're a feature, not a product.

Excel was a vastly superior product to the competition. I used it back when Lotus was still running DOS character mode. The only really competitive product today is Google Sheets, because collaboration is a killer feature. And Google Sheets is doing very well.

Java was never a Windows replacement. And as a platform it is doing just fine.

Instant messaging died out due to Facebook/Snapchat/Twitter/Slack/etc. It's not like everyone is using Skype (which MS bought, not built, and long after the IM dust had settled).

Google dominates email, not MS/hotmail. And spammers killed the "run your own email server" approach. It requires professional knowledge to get email delivered these days.

None of those examples have anything to do with "embrace, extend, extinguish"?

They all do. IE is eee of HTML and the web. MS Java and VisualJ or whatever it was called was eee of Sun Java. ActiveX was eee of the web browser.

That wasn't Microsoft's choice, that's just de-facto reality

That's what BG said in his deposition, but MS was the only company who embedded IE into the OS to make ActiveDesktop and put VBScript and ActiveX into IE.

Java was never a Windows replacement.

AWT and Swing and browser applets were Windows replacements.

Instant messaging died out due to Facebook/Snapchat/Twitter/Slack/etc.

You're talking about things that happened ten years later. There was a time when one app would connect to every network.

Google dominates email, not MS/hotmail.

MS Exchange dominates email.

>They all do. IE is eee of HTML and the web. MS Java and VisualJ or whatever it was called was eee of Sun Java. ActiveX was eee of the web browser.

All of those efforts failed miserably. They have nothing to do with the actual reasons why Netscape died, why Firefox is trending down, why Java never became dominant, why Open/LibreOffice never replaced MS Office, why open chat protocols were replaced by chat services from Google, Facebook, Discord and Slack, why Java applets never caught on, etc.

"MS Java and VisualJ or whatever it was called was eee of Sun Java."

You mean C#?

>Steve Jobs put it best (to a different company): You're a feature, not a product.

Dropbox, which has also been falling apart trying to become a product.

> Netscape died because browsers became a part of the baseline operating system

It's hard to decouple the fate of Netscape from their catastrophic decision to rewrite their entire codebase from scratch:

https://www.joelonsoftware.com/2000/04/06/things-you-should-...

That was the final nail in the coffin, but they had been pushed aside before that. Hence the final "hail mary."
> Netscape died because browsers became a part of the baseline operating system.

You're ignoring the actual EEE part.

Microsoft didn't just include a browser, they included a browser that didn't follow standards and had a bunch of its own extensions. Then since Windows was the largest platform, lots of web pages started using Windows-specific ActiveX controls and other IE-specific features on websites, and users had to switch to IE even if they preferred a different browser.

> Steve Jobs put it best (to a different company): You're a feature, not a product.

Steve Jobs did the same thing. You have to use their browser on iOS and its purpose is to be less capable than competing browsers to force developers into making native apps where Apple gets 30% of the developer's revenue and can exclude apps that compete with their own services.

> Excel was a vastly superior product to the competition. I used it back when Lotus was still running DOS character mode.

The history of Office goes like this. There were many competitors and many users preferred the competitors, but Microsoft intentionally made them crash on Microsoft operating systems so that people would use Office instead.

Then, once Office had the most market share and everybody was locked into it because all their documents were in its proprietary format, they used the revenues they denied to competitors to make Office better. Now you say, look how good it is! But how did it get there and why is nobody else?

> The only really competitive product today is Google Sheets, because collaboration is a killer feature. And Google Sheets is doing very well.

now their competitors are only the likes of Google and Facebook who get where they are by leveraging their own dominant market positions in other markets.

> Java was never a Windows replacement.

It wasn't supposed to be an operating system. It was supposed to make it easier for users to switch operating systems, which Microsoft successfully impaired.

> Instant messaging died out due to Facebook/Snapchat/Twitter/Slack/etc.

This is the one where they basically failed, because the network effect counterbalanced the leverage of the Windows monopoly. If your friends had ICQ then you installed ICQ even if you already had MSN Messenger installed as part of Windows.

But it was clearly still an attempt to do EEE. If they'd succeeded in getting MSN Messenger into a dominant market position then they could discontinue or cripple the non-Windows clients and lock people into Windows with it.

> Google dominates email, not MS/hotmail.

Outlook.com has 400 million users.

> And spammers killed the "run your own email server" approach.

Large email providers killed the "run your own email server" approach, by marking email from small email servers as spam even when it wasn't. Plausibly on purpose.

And we got large email providers to begin with because they were the ones who could overcome the Microsoft lock in to Outlook/Exchange. Gmail did that by offering 1GB free storage back when that was expensive and subsidizing it with Google Search revenues.

> You have to use their browser on iOS and its purpose is to be less capable than competing browsers

This is not a serious take. Safari on iOS is a very capable browser and crazy things have been done with it. Where it does make it difficult to replace a dedicated app, this can be ascribed to security more easily than "Apple wants one of its major iOS features to be bad".

> Microsoft intentionally made them crash on Microsoft operating systems

This is not a serious take. I'd love to see proof of this. There are lots of reasons Office became dominant, some of them even anticompetitive; "MS made competitors crash" is probably not one of them.

> It was supposed to make it easier for users to switch operating systems

This is not a serious take. Java was never intended to make it easy for users to switch operating systems. Sun did not make "supplant Windows!" one of its KPIs, and the continued dominance of Windows is neither here nor there when evaluating whether Java was successful. Java's pitch was to make writing platform-independent code easier, but at best that's tangentially related to having users switch OS's.

> Outlook.com has 400 million users.

So?

> And we got large email providers to begin with because they were the ones who could overcome the Microsoft lock in to Outlook/Exchange.

You're replying to a comment about people who run their own web server, so talking about Exchange lock-in is neither here nor there.

> lots of web pages started using Windows-specific ActiveX controls

Ah, that explains why ActiveX took over the web and why I'm forced to read HN on IE. /s

> The history of Office goes like this.

I'm guessing you are way too young to have experienced it, because the history of Office was nothing like that.

> Java

Again, I'm guessing you didn't actually live through that era. I did. Hell, I even wrote Java desktop apps for a living in the late 90s. Microsoft did absolutely nothing to prevent Java from taking over the desktop; Sun managed to accomplish that all on their own.

> But it was clearly still an attempt to do EEE.

You can talk about "attempt" all you want, but there are still zero examples of EEE being successful.

> Outlook.com has 400 million users.

"Gmail is a free email service provided by Google. As of 2019, it had 1.5 billion active users worldwide." - Wikipedia

> Large email providers killed the "run your own email server" approach

Man where do you get this stuff? I wrote a mailing list server that had a brief moment of popularity, and the underlying (Java!) smtp library that lots of other folks use still today. I know a thing or two about smtp, and I gave up running my own email servers a long time ago. Spam fighting requires a massive engineering team. Barring some sort of massive change in the protocols, the home email server is dead dead dead. Microsoft and Google are the symptom, not the cause.

>Fun fact: every single example in the Wikipedia "embrace, extend, and extinguish" ... is something that failed

And probably wasn't GPL-licensed. That part introduces a slight problem to the EEE process. ;p

Not sure why you are being downvoted, you have a point. Though it'snot about license, but more about market position. I'm also afraid that MS still has plans for Linux, but they are on the first E atm.
I have exactly the same fear. Trying to lock down boot on PC is pretty much exactly what they seem to be doing. Hello to shitty OS on PCs... (yes, every mobile OS is pretty bad)
I think this is something MS could try.

However the whole computing paradigm has changed. Laptops/PCs are now 'trucks' and the 'cars' of the computing world are phones, tablets and consoles. Ie. Laptops and PCs are mainly used for work now, and lots of that work is being done on Linux. Lenovo ThinkPads can be shipped with RHEL, Fedora or Ubuntu. Several Dell laptops (including the highest end workstations) can be shipped with Ubuntu. Framework and System76 of course are Linux friendly/default. And most other makers allow disabling UEFI from the BIOS menu.

Considering how much of MS' value is derived from Azure and that their main competitors are Linux-first (Amazon, Google), it'd be idiotic for them to try kill Linux again. You're right though, they'll probably try. I just have faith that enough hardware makers care about their users more than propping up WinTel like the old days (both Linux and AMD have made massive inroads).

I wouldn't put anything pass these guys (MS). Don't think you are overly paranoid. With MS, they backdoors, one can NEVER be too paranoid.
Those fears will be reasonable when most computers will not allow disabling Secure boot. Right now that's not the case for x86.
It's perfectly reasonable to be skeptical of the measures being put in place which would make this possible, especially given MS's history.

To give an analogy, there might have been a time in East Berlin where people could have said: "It's fine that they check papers before crossing into West Germany. Your fears will be reasonable when they no longer permit entry."

Secure Boot is about to become the consumer standard with Windows 11. Popular online games are already starting to require it. I wouldn't be surprised to see Netflix and other streaming services require it.
Keep in mind that this is currently possible only because some governments were looking into persecuting MS for abuse of market power when Secure Boot was standardized.
It doesn't matter if that's the explicit goal. Given the opportunity MS would do it.
Yeah, I think you're being overly paranoid.
> Instead of open-sourcing Windows and treating it as open technology something the Cloud group would love they still jealously guard it as proprietary.

Ignoring the question how much revenue they'd lose from such a move I don't think they could easily.

Windows doesn't consist only of Microsoft's code, but there is the whole legacy with OS/2, DEC/VMS and certainly a bunch of external companies contributing device drivers and things. I assume there are notable parts where Microsoft can't easily figure this out anymore. Code was moved and changed over time and for open sourcing they have to be sure about each line being ok to open source.

Look at Sun's process for opensourcing Solaris and where even OpenSolaris still had some binary blobs in some parts in the end after they went through the multi year process of going through it and rewriting different parts.

There are legacy parts, but OS/2 subsys was never used, VMS was never supported. The OS has become more modular, would be easy to deprecate ancient stuff.

If anyone has the resources, it is MSFT, however investment in Windows has been on a downward trend, so don't hold your breath.

Who knows what's in the contracts ... and what code moved around and survived where MSFT has no license to share code. Sure most of the things are out, but one helper function here, one there, maybe some data type they are still using relies on code from there ... it is a huge codebase grown in different ways over decades. And a few lines in an old subsystem and some and some successor of some original license holder can use it to refinance all their investments by sueing Microsoft.
Microsoft still makes billions from selling WIndows OEM licenses. According to their financial report such sales increased 9% from 2019 to 2020. It appears you can absolutely make tons of money selling desktop operating systems in 2021. I'd be delighted to make even 1% of Microsoft's Windows desktop OS sales.

Personally, I think exactly what is ruining Windows as a desktop OS is trying to cloud-ify everything and render it more like a "future of computing" operating system packed with useless AI, Metro apps, cloud integration, app stores, and locking it down for normal consumers. I mean, look at Microsoft deprecating fantastic apps like Windows Backup or Windows Snipping Tool. They want to transform Windows into a "hip" and "cool" Nadella-vision OS and that vision sucks.

Microsoft is a fractal of infighting and groups at war with each other.

Cloud Services are cool, and Windows are the weird ones. But within Windows there's the kernel team who are cool and customer-focused, and the User layer as the weird ones. Go another level deeper and you have fights between C++ and .Net, or historically the crusade for COM+. Or the transition to UWP, which clearly at most half the Windows UI team got behind.

Obligatory web comic: https://www.globalnerdy.com/2011/07/03/org-charts-of-the-big...

It’s true that there are cool groups within the organization but I don’t trust the people above who can guide via budgeting.
> Instead of open-sourcing Windows and treating it as open technology something the Cloud group would love they still jealously guard it as proprietary.

I agree with the rest of your comment, but I don't think this really works. Windows remaining proprietary is something that makes sense from MS's POV. If, say, Windows 12 were to become open source, would you use it as a development platform or as a server? I for one wouldn't, the reason why developers don't want to be on Windows isn't that Windows costs money, it's that Windows is bad.

It only makes sense for MS to continue milking Windows until the bitter end. Eventually it won't be worth the effort any longer and Windows 19 or something will be a Linux distro.

The TPM thing for Windows 11 makes all my home PCs (5+) not meeting the requirements for upgrade, even though they all work totally fine. I am seriously considering moving to Linux desktop and I am doing research on it. I have tried a couple times in the past and my experience with Linux desktop was terrible. I am still not convinced to make the move. Maybe I need a Chromebook kind of OS for my family and a different one for myself.
You might have heard it before, but here it is again: your "terrible" experience might be (partly or not) due to the fact that you've been using Windows for decades and know how to solve problems. On desktop linux, you might not, so you experience friction and frustration.

Desktop Linux is indeed somewhat behind OS X or Windows. I'm currently on (somewhat dated) Ubuntu 18.4 from 2018, and it is more or less on par with Windows 7 from 2009. For me that's not a problem, I liked Windows 7 and actually stayed on it until 2018, but keep this lag in mind.

I gave Windows another chance with Windows 8 after switching to Linux Mint KDE of all things and that finally killed off any desire to use it as a daily driver for good. I found KDE very understandable coming from Windows, in all honestly it felt more like "proper Windows" than the industrial accident that was Windows 8's UI ever managed.

A user-friendly Linux distro like that was a gateway drug into becoming a terminal jockey though.

It’s definitely Linux that’s the issue. Too many driver issues, software version issues, lack of unified distro environment.
Yeah, I feel your pain. I hate Windows more and more every day, but whenever I try to use a Linux Desktop for anything serious I find I hate that even more. Some day Microsoft will achieve a level of awfulness that makes Linux Desktop's awfulness the more tolerable alternative for me, but it won't be today.

Maybe if I get really really lucky Haiku will be usable for my use cases by then, but that doesn't look very likely.

I just can't bring myself to use Windows anymore, but I agree that Linux Desktop just isn't there. I'm mostly on Macs these days, but I still have a Dell laptop running Fedora 35.

Every other day something breaks. We're spanning the entire range here, from petty stuff like Firefox missing fonts, to straight up the kernel not booting. And obviously everyting in between, like apparently the only way to receive Exchange mail is to pay for a Thunderbird add-on. Or setting up the VPN requiring arcane SELinux magic.

I honestly don't see how one can recommend desktop linux to someone without at least a minimal knowledge of how the OS works and some decent terminal-fu.

As sad as it is your parent is right. Non-technical users should just get a chromebook.

I'd never recommend Fedora, non-LTS Ubuntu, or Debian Unstable for a Linux newcomer. Any time you run a development release, you risk breaking something whenever you update.

I've been happy running Cinnamon on Ubuntu LTS releases, and I like what I see playing with KDE as well. I've never had a kernel fail to boot (other than by hardware failure), and AppArmor is, IMO, far better at staying out of the way than SELinux. I can't help on the Exchange stuff, I've never had to deal with that.

It would be a little easier to just run Mint instead (Cinnamon being native there), but on Ubuntu you can just "apt install cinnamon-desktop-environment" and then choose Cinnamon the next time you log in.

Exchange access outside Outlook and Apple's Mail.app is PITA anywhere. Part of why you need to pay for Thunderbird add-on is that Microsoft asks for license fees. Nobody is going to pay that for you (and for Apple, it is hidden in the overall package price).

What SELinux magic you had to do for VPN? All VPNs that I've set up were fine with just clickety-click in the GUI.

Well you are running Fedora which is a bleeding edge distro aimed at Linux devs and is a testing ground for red hat. You’d have a lot less problems on Ubuntu.
Not really; I've never had any problems with Fedora. It is actually pretty polished distro.
I didn’t have many problems either except in regards to my nvidia graphics card. I’ve used Ubuntu for about a year (coming from Fedora for 3) and Ubuntu is more stable in that regard.
I hear this a lot. What I don't understand is why my own experience is so different. I don't find Linux desktop awful at all. On the contrary, I find it to be the best (by far) experience as compared to both Windows and MacOS. It's so much more convenient. MacOS is kind of OK, but Windows is nearly unusable.
Probably a lot of it is just that we use personal computers differently and have different expectations of them.
What I've seen it is not just expectations, but also experience/inertia.

When somebody is used to (for example) Windows, he then approaches any problem the Windows way. But the Windows way may be not the optimal one on any other system. But since it was optimal on Windows and the person doesn't know anything else, then he will see suboptimal results anywhere else.

It is not just Windows -> Linux; the same problems will be experienced with Windows -> macOS. Even Microsoft had the same problem, when they redesigned Windows elements and the uproar it caused.

The key to enjoying Desktop Linux is giving up on the desktop environments that try to ape the Windows/Mac stuff and just biting the bullet to a tiling WM. It's a steeper onramp (e.g. you somehow have to know you need rofi to have an alt-tab window switcher) but it leads you to something different and better.
More generally, I only find desktop Linux tolerable when I start from very little and build up to what I want, rather than starting with a full DE—even the heavier "kitchen sink" XFCE variants are too heavy to start with.

There was a period of maybe three years in the late '00s when Ubuntu was really killing it with their defaults & configs under Gnome2 and it really was quite nice, but before and after that, I've found going minimal and all hand-installed to be the only way not to constantly be angry at my machine and experiencing all kinds of mysterious brokenness.

In fact, accepting that what you can productively and pleasantly do with desktop Linux is a different (though overlapping) set of things than what's productive and pleasant under Windows or MacOS, and not trying to do those other things at all, is the path to contentment under desktop Linux, I think. At least until it gets a lot better, which, with all the Wayland shake-ups, seems really far off, if it ever happens. I think it's more likely Google's replacement will get actually-seriously-no-bullshit-quite-good for desktop use, before Linux does.

For the most part, though, I just stay on MacOS these days. All the time I lost to Linux during my decade or so of using it as my main OS on desktops and laptops taught me a lot, so wasn't completely wasted, but I just can't be bothered anymore. I just want to start up the machine and get stuff done. MacOS gives me plenty of glitches, I don't need even more, plus a ton more crashes and weird behavior. And I do want my computer to just do a bunch of nice stuff for me automatically, without constantly (instead, merely often) breaking or glitching or requiring me to set all that up.

Maybe look at using Android for the desktop. Example Android-x86 (https://www.android-x86.org/) and others. The thing is, many people are already using Android for their smartphone. The jump to laptop or desktop is arguably smaller than many make it out to be.
Don't consider - just do it. One at a time. The more you consider, the lesser the chance that you actually will.

Linux desktop is terrible only as long as you perceive it as an "other Windows" thing. No, it is something completely different. Instead - make a commitment, embrace it - and you will learn to love it.

I actually tried this approach before, but eventually something would be come a last straw to force me give up. For example, as I know, there is nothing comes close to Windows RDP on Linux Desktop. On Linux desktop, it's very hard to remote to the local desktop session from remote, also no automatic resolution adjustment like RDP.
I don't know if I buy this explanation to be honest. I guess main business is office because nearly every company on the planet uses it (for no good reason in that vast majority of cases, and please don't be the guy that defends 'cloud-like' auto-save).

Azure? Not too much touching yet, I am an AWS guy (I can complain here just as much). But their office software and worse their office software APIs are abysmal. I am not registering an app and describe its function to access their beta graph API just to add a task to Planner. I fired the customer that wanted that.

Desktop does make money since most people use their office suite on a Windows PC. Windows isn't that important anymore, but it is still present. And the direction they go with TPM and notebook camera requirements is predictably bad.

You cannot open-source Windows anymore, that would probably be a legal nightmare. Maybe parts of it of course.

Some newer parts of Windows are open source, and in other places they've added open source software to the overall distribution (such as OpenSSH). But yeah, almost nothing legacy in Windows has been or will be open sourced, due to the difficulty of tracing which lines Microsoft actually owns and which ones they only have a license to.

For legacy components that have been open sourced, the only one I can think of is conhost, which they're trying to get rid of anyway in favour of the new (and fully open-source) Windows Terminal and its console component OpenConsole.

For what it’s worth, OpenConsole is conhost, and any improvements we make there go right back into Windows. We’ll probably[1] never get rid of conhost; the worst we’ll do is make it possible to update it without updating Windows so that folks don’t have to wait so long to get bug fixes.

[1]: so long as I am the engineering lead. You know how these big companies are :|

(edit to add: This isn’t meant to detract from your point, just to add additional info.)

I could have sworn that OpenConsole is a complete rewrite of conhost. Having double-checked the terminal project's readme, though, I see that I was mistaken in this belief.
Do you have any plan for Terminal to be able to be the default console in Windows 10, just like it can in Windows 11?
Where do you place the teams in charge of the steaming pile of manure that is Teams, SharePoint, Dynamics 365, Office 365 et al? Seemingly these turds have also dropped from the cloud faction. My experience seems to be that excluding dev tooling, all attempts by MS to bring productivity tools to this brave new Internet thing have failed dismally.

Office 365 vs Google Suite is not even funny, if you have to use O365, that is. Dynamics 365 looks fancy at first, but make a few clicks towards settings and you can witness 5 different UX languages, one half-assed modernisation attempt laid on top of another. Plus the churn from ongoing attempts, pointless rebrands and nonsense invented by bizdev guys. Add the famous MS back compat policy and you have a special kind of hell.

Under Satya Nadella watch they brought in these changes. Windows 10/11 are dark pattern products.