Ask HN: Future of Windows operating system and application | Hacker News Reader

Ask HN: Future of Windows operating system and application

14 pointsthorin18 comments

Now that windows is more tightly integrating with linux and supporting wsl etc and applications are increasingly xplatform e.g. vs code in electron on linux, apple etc where do we see this ending. Is there much change in how windows is evolving under the hood or in the way (previously) native applications are being developed? Although Windows is seen as less relevant these days mainly due to the increase in share of mobile and tablets, in the corporate world I still see everything I do with everything client is 99% windows in terms of the desktop/laptop infrastructure (even though it doesn't need to be).

Are you aware of any current writings on the subject and strategy of the Windows OS team?

View on HN

Linux is not integrated to Windows. WSL is just a VM with some convenience features.

The entire PC and Windows ecosystem is based on that the purchaser is not the user. Corporate will keep purchasing sub-par hardware with a sub-par operating system and sub-par software, because they're looking at a primitive spec sheet and feature list, without taking anything else into account.

Developers (for corporate customers) will keep making their software as cheaply as possible and the user experience will continue to decline while Windows and PC sales increase to corporate customers. These platforms will continue to decline in relevance for consumer use.

That's my prediction.

Vscode had been wildly, and widely successful for Microsoft.

I think that's the direction Office is going. I think that model will be used on other applications...

Personally, I keep thinking of switching to Linux because the performance (compile) is higher, and so much code is easier to build and use.

I would be beyond thrilled if Microsoft would allow users to uninstall/decrapify Windows. The hypocrisy of - the browser is part of the OS and cannot be uninstalled and then it was when edge was released is insulting to users.

Since Channel 9 died, I'm not sure there is any good source for insight into what people are working on in terms of OS internals.

You can look at a list of new features for Windows 11, but aside from your mentioned support for Linux/Android, there isn't anything that would impact how most apps are developed: https://en.wikipedia.org/wiki/Features_new_to_Windows_11

People to follow to keep up on what Microsoft people are working on:

Scott Hanselman (https://bsky.app/profile/scott.hanselman.com)

Jen Gentleman (https://bsky.app/profile/jenmsft.bsky.social)

Kayla Cinnamon (https://bsky.app/profile/kaylacinnamon.bsky.social/)

Mark Russinovich (https://bsky.app/profile/markrussinovich.bsky.social more Azure these days, but he still posts about user stuff too)

Clint Ritkas (https://bsky.app/profile/clintrutkas.bsky.social)

Craig Loewen (https://bsky.app/profile/craigloewen.bsky.social) WSL guy

They seem to post on BlueSky mostly, so I've linked those.

In most enterprise jobs aside from specialized, vertical markets, the operating system doesn’t matter. So why not use Windows PCs? They are cheaper, companies like Dell have great enterprise support contracts, IT installed malware (ie MDM software) is mature, etc.

In tech oriented companies, employees are usually given a choice between Windows or Macs.

I agree for the most part, however the question is more about where the windows product itself is moving.

Nowhere as far as anything innovative or ground breaking (the same for Macs). Because of the enterprise lock in, Microsoft can’t make drastic changes to Windows.

If anything, they are still trying to get away from their x86 dependence and move people to ARM. But on the other hand, from what I heard about the latest surface laptops running x86, they get great battery life and are only slightly worse in performance than ARM equivalence.

As far as software, only vertical markets where software has to be done offline or performance is critical will we see any Windows native software.

Windows isn’t even Microsoft’s core concern anymore. It’s Office. Server products and Azure.

The windows product itself is being enshittified into an ad delivery machine, and surveillance tool.

As for writing apps for it: use dotNET, IT work will be using powershell

Most large well run businesses use MDM and have rules to preveht information leakage to Microsoft and ads.

When I use to buy Windows computers, I always bought the business line. Are they just as bad as the consumer products these days? The last time I bought a Windows PC was 2016.

> So why not use Windows PCs?

If "the operating system doesn't matter", there's this thing called Linux. You can still buy Dell computers and have their enterprise support, and run RedHat or whatever, and have their enterprise support.

So, other than habit, why should Windows be the default?

Why use Linux? Everyone knows Windows, it only adds $30 to the cost, it’s easy to hire cheap Windows admins, etc.

I was saying that the operating system doesn’t matter between Windows and Macs where there is mainstream commercial software - especially Office365 - and people are already comfortable with it.

Office 365 apps don't work on Linix so if your company uses it, operating system does matter.

This may sound unexpected, but the Win32 API is still perfectly functional, and is probably your most solid foundation for any native development efforts you want to ensure continue to work long term. It's found a degree of support on other platforms (e.g. Wine/Proton, even quirky upstarts like ReactOS) and if Windows goes the way of the dodo my gut says the industry will find a way to breathe life support to that layer. As a bonus your binaries can be small and self-contained. .NET looks like its future is bright as well (was really glad to see Core, open sourcing, and their strides to unleash the stack beyond Windows).

No inside info here, but as a lifelong Windows developer and user, I'm disappointed in the route Microsoft took with their OS. Feels like decades of poor business decisions and squandered opportunity. Throwing all kinds of new methodologies out there without a clear vision. Arbitrarily designating older technologies as outdated then doing a poor job of reinventing them (looking at you WinFS), and an obsession with trying to add "more" when cleaning up and refining what they had first would have served better. I'm still annoyed by the whole "we need to fatten all our directory names to show off long filename support" which to this day left us ugly artifacts like "Program Files (x86)", linked folders you can't easily open, redundant places for registry settings (...don't forget SysWOW64), autoruns spattered all over the place (the SysInternals tool of the same name does a great job of surfacing), the list goes on. The transition from 16-bit to 32-bit was cleaner than 32-bit to 64-bit, I suspect because the developers who worked on the former were "closer to the metal" and could just make the changes in the code instead of spending 10X the effort on workarounds. If there are insiders in the room who say otherwise I'd love to be corrected.

Windows 7 was the last edition I could tolerate. The user-hostility of Win 10 and 11 is a non-starter for me (ads, telemetry, ramming broken updates down your throat that brick your PC as the worst time, their trajectory toward forcing TPM so that you no longer own the keys to your castle, etc). They're copy-catting some of the worst practices by competitors in a big race to the bottom. It's a real shame, because the kernel team made some great improvements under the hood in the meantime.

If this sounds like an old person rant, I guess it's because it is. It hurts, as I grew up with this OS, became incredibly proficient with it, gathered and built an ecosystem of tools to support how I work, started a business on their stack, helped customers adopt it to be successful, cultivated a reputation, etc. I still hope someone with authority at Microsoft smartens up and sets a course correction, the way they were late but eventually did embrace the Web and open source. I was always skeptical of the whole "Windows 10 is the last version that will ever be" nonsense, and waited it out hoping when they readopted version numbers Windows 11 would arrive along with a renewed focus on putting the user first. But at the moment I'm resigned to the likelihood I'll have to figure out which version of Linux will work best for me and get serious about adopting it.

This is an excellent writeup. Even though I come with a bias against MS, I did some work on the .NET framework and was just blown away with how great the dev experience is. Maybe I’m overselling it but I wouldn’t consider any other platform (except that it’s managed by MS). The old school win32 api is surprisingly powerful too.

Then looking at the spider web of new technologies they’ve released, half-released, and half-deprecated it makes me seriously doubt the future of MS. I mean they’ll always be a profitable, popular company just not whose tools I want to be locked into.

Working in the SharePoint ecosystem, for example, was totally miserable and I’ll never do it again if I can avoid it. Still love C# though.

Luckily .NET is not married to Windows and MS ecosystems at all for good 9 years already. I'd be long gone and using something else if it did :) (very happy I've stayed though)

Agree with your worries and hope they find their engineering excellence again, because while DevDiv teams working on .NET are isolated from problematic influence, it would be a shame for MS troubles to affect them. It's just too big of a project to continue without significant funding and there is probably not enough community goodwill and traction in absolute numbers (which are wasted on worse languages like Go) to carry the torch.

MSFT seem to be paying more attention to the user experience of Office than they do of Windows.

They do have a small number of initiatives, e.g. WinRT (a replacement for Win32), and they're introducing Apple-style sandboxing for apps distributed using MSIX, their "new" packaging technology. However, this tech is hard to use out-of-the-box. I have a company that sells a cross-platform tool which makes shipping code that uses new technologies for cross-platform apps much easier [1], and in particular, lets you ship from non-Windows platforms like Linux CI machines or Mac developer laptops.

There's DirectStorage, a new API for games developers which I think is a port of tech developed for the Xbox.

But otherwise, not really no. Windows appears to be in maintenance mode. Large parts of its development were outsourced to India, the higher level executives are disengaged. Their social media is full of left-wing Current Thing viral sharing rather than anything about developing Windows apps (see the bsky links in another post for examples). They are in some core tech areas nearly 20 years behind Apple. If there's any kind of strategy, it's not widely known. They seem to be faced with two intersecting problems:

1. Nobody targets the Windows API anymore directly, the only app that matters is Chrome and if the Chrome guys don't add support for something then that API will just go unused. Browsers have a policy of not exposing tech that's platform specific. So this dynamic is killing all desktop operating systems, not just Windows. Nobody is investing much in desktop Linux anymore, Apple do still invest in macOS albeit not at a huge level, I think because there's still an ecosystem of Apple-native apps out there. But even so they've focused on initiatives like Catalyst. Microsoft have tried to fix this by forking Chrome into Edge, but the Edge team seem reluctant to add any unique Windows features, perhaps because the Chrome codebase just isn't set up to handle that kind of extension.

2. The Windows codebase has reached end of life, IMO. It's notable that the new APIs or changes they make to existing APIs are all very buggy or have obvious design problems, and that Microsoft's own stable of apps don't get rewritten on top of them at any kind of speed. The core Win32 API is what nearly everything uses but was abandoned years ago, and if you use it your app will by default still look much like a Windows 95 app would. They seem unable to change it in any meaningful way, and also unable to design a sufficiently compelling competitor to HTML.

It's still amazing to me that Marc Andreessen said in the mid 90s, "Netscape will reduce Windows to a poorly debugged set of device drivers" and that prediction actually came true.

[1] https://hydraulic.dev/

Reply on news.ycombinator.com