PC gaming just shook a little. I have decades old games that still work because Microsoft still supports Windows. uTorrent entirely exists because of Microsoft's commitment to break-your-back-backwards-compatability in Windows. DirectX is 25+ years of of PC gaming's history.
That could disappear like a console generation. That's a little disconcerting.
The reason uTorrent can stay small is because it exploits libraries that have existed in Windows since NT, mostly for GUI work. There was also development that incorporated the HTML Application library [1]. At least that was the case when I worked on it back in 2014.
I am specifically referencing the original PC version. There's a whole other library in the code dedicated to the mac platform implementation, as well as Linux. The reason uTorrent on Windows started small was because it didn't include much of any static linkages, but this depended on dynamically linking some libraries in Windows that Microsoft never removed. This is the dedication to backwards-compatibility I refer to.
On the contrary, I think perhaps the best thing for the preservation of the history of windows gaming is for windows to stop changing. A frozen or deprecated win32 API will give WINE (which is already quite good) breathing room to catch up and fill in the gaps, without having to continuously cope with the shifting sand of the latest experiments with windows. We'll be able to run old windows games forever, rather than just as long as Microsoft deigns to allow it. DOS games are fairly well preserved at this point for similar reasons.
Of course, then we'll have to deal with whatever the next platform is, but that's just the reality of the industry.
I understand where you are coming from. You are correct that an constantly changing desktop experience is sub-optimal.
I fear we will lose the Microsoft as the silent steward of backwards compatibility, DirectX support, getting manufacturers to make great hardware drivers, the list goes on. It's this part of Microsoft that Windows will lose, because it isn't customer facing.
I think the fallout is exactly the opposite of what you want. A de-emphasized Windows means that the customer experience stuff is pushed forward while the above suffers. The only bulwark against that is how deeply invested most enterprise companies are in the Windows universe, driving many of those behind-the-scenes fixes and changes.
> Of course, then we'll have to deal with whatever the next platform is, but that's just the reality of the industry.
Gaming on the PC was, in many was, a repudiation of that idea. We could have one continuous open system that ran everything from old games to new games. That's part of the point the emulation crowd wants to make as well. We don't have to live with platform churn.
But I'll be the first to admit Stallman was right. We collectively ceded this power to a private corporation, and we could lose much of what we had because we weren't good stewards.
Actually... It's more common than not for an old game to not work correctly (or at all) in recent versions of Windows without hacks and compatibility wrappers. Try to run a DirectX 6 game on Windows 10. Today, WINE is closer to be that "one continuous system" than Windows is.
A "continuous platform" was more aspirational than fully realized. I bet I could find some of my DirectX9-era games of mine that won't run past Windows 8. But the lack of an hard artificial churn like consoles combined with an open continuous API made PC gaming feel resilient in the face of the chaos.
Yeah, my Rocksmith 2014 doesn't work at all in Windows 10 due to some weird compatibility thing with the cable that Ubisoft refuses to acknowledge. Meanwhile it runs great in Wine, thankfully.
My hope is that it will put more effort into projects like ReactOS, which will get us a binary-compatible version of Windows that is free and open-source.
Microsoft in the last few years had a see-the-light moment with Open Source. If the stars align and I'm good this year, Microsoft might release some core Windows and DirectX code to the community.
Because MS is pushing more and more consolized and controlled setup. They want Xbox like environment on PC (UWP). So if want true PC gaming where you can mod what you want and how you want, Linux is the best bet.
The only evidence I see of Microsoft trying to make PC gaming more like console gaming is them selling games through the Microsoft Store (I think that's what it's called now) and including the Game Mode stuff in Windows 10.
The vast majority of gamers using Windows buy their games through Steam, Origin, or other stores that Microsoft has nothing to do with. Unless we see a massive market shift then I don't see this having any affect on the future of PC gaming.
As for the game mode stuff, it's easy to turn off if you don't want it so I don't see any issue with it.
I also think that most gamers (myself included) would be okay with PC gaming becoming more consolized.
Back in the 90's you had to mess with configuration files in both the OS and the game you were playing far more than you do today. In a sense, it's because PC games have become more consolized - they "just work" most of the time. People like that. I hope the future of all forms of computing "just work" better than they do now.
I will concede that if you want absolute freedom of customization, Linux is the best OS for you even if you don't plan on gaming at all. However, saying that Linux is the future of gaming is about as accurate as all of the claims that "20XX will be the year of the Linux desktop".
I had to install some wine libraries for Arcanum because current Windows doesn't support DirectDraw properly. I imagine it will only get worse, but open source will fix that.
Correct me if I am wrong, but it doesn't have to be inside cmd. A lot of people use cmder/conemu, and it works great with WSL. You can even put your own shell on top of WSL (zsh, fish, etc.), if you are inclined to do so.
I tried insider for a few months and found the frequent updates too much, have some stability issues and stop it. Maybe it is a good reason to to start it again.
I think you mean conhost (cmd is the shell program, not the terminal). It does have it's rough edges but it's much improved from where it was only a few years ago and its far from the only way to host a WSL session. This includes support for ANSI escape codes (TERM=xterm works fine) and better colors (though you may need to reset the defaults if you're upgrading from an old system). Lastly, there are plenty of other terminal apps which make this much more familiar for folks who prefer things like tabbed windows, or theming.
A second point, WSL programs don't rely on an active conhost shell to keep running. You can kick them off and run them in the background just fine. On one of my machines, I have a full i3 session running with no conhost window in sight. Everything works seamlessly, I can talk to my editor running as a Windows app, cut & paste back and forth, see file change notifications, hook up my docker command with my docker for windows VM, and so on. It's nice enough that I miss it when on my Mac and get annoyed with some of the fussy application alternatives on the typical Linux desktop.
I use a native linux box for most things, but I do have a windows PC for time wasting stuff. WSL is very close to being a good enough for most of my uses. It is an impressive product.
How does ConEmu compare to iTerm2? Why do you think its not up to the mark? I have recently moved to a company using windows computers and ConEmu with WSL has been my lifeboat.
Hey there, I'm actually on the windows console dev team. Any specific features you're looking for from iTerm2? There's already hundred of items on the backlog, but I'm always curious what things in particular people are looking for.
Yes amazing because for the best part of 20 years, nothing significant has changed and now it’s functional enough to get work done.
And I don’t care about the *nix implementations because the first thing I do is start tmux and use that. Same as I only care for the Linux userland which is what WSL gives me.
posh-git is a great powershell module that shows your current branch, has customizable colors, and adds tab completion to git commands. I think that stuff should stay as add-ons rather than be built into the console host.
Auto updating the prompt isn't possible in CMD unfortunately. I do believe it's possible in powershell, but that's not my cup of tea. I do have a batch script that gets the branch and re-sets my prompt. You could create a git.bat that calls such a script on every git command, or like me, create `vf.bat` that I use to cd and update the prompt at once.
Also in this picture: cmd.exe running inside tmux in WSL. Coming soon to Insider's Builds near you
I'm interested to understand the delta here too. I used cmder daily for work and iterm2 at home and I am quite happy with both. Sure they are different, but they are always going to be.
Makes sense to me. They are ahead of AWS around K8S support. That kind of thing, if they do it in multiple areas, will erode Amazon's current lead. Azure is also noteably less expensive than AWS. They are already eating into Bezos' market share.
Also the whole office 365 plan gets F500 companies already tied into Azure. It's sort of like the drug dealer's "first one's free" tactic. They will, for example, spend money on direct connect network for O365, which makes Azure immediately more attractive. And their Active Directory info is necessarily already federated to Azure.
AWS is in a really weird place right now. They're powerful and have a bright future, obviously. But it feels like its already become a legacy platform in the cloud world.
Look at what they've released in the past two years... their AI stuff is second rate compared to Google, and seems like its just there to check boxes. Their serverless stuff beyond Lambda is overpriced, like Aurora Serverless and Fargate. They're dragging their feet on EKS; it was announced 5 months ago, we still don't have a release date, and the latest webcast (last week) they showed made it seem either really bad or really unfinished, like you have to involve CloudFormation to provision instances, manually manage VPC subnets, etc.
Its a world where they have a core set of "legacy" building blocks that are really good and well-tested (EC2, RDS, S3, SNS, SQS, a few others) but if you go beyond those you enter uncharted territory that mostly exists to seem hip and check boxes.
GCP, on the other hand, has legitimately powerful products at higher abstraction levels like GKE, Firebase, DialogFlow, and their AI stuff. I know less about Azure, but they seem to be doing well checking boxes and pushing toward more higher abstraction levels with things like AKS, service fabric, etc. Where's the actionable innovation at AWS?
Agreed. Looking forward to when Azure, GCP, and AWS have more equal marketshare. That will fix some current cloud shortcomings. Like stupid high egress rates. Or Amazon's lack of a real Spanner competitor.
Though, I get some of the egress upcharge in GCP. Their global network is worth some premium. Just not the current predatory pricing.
To be honest it is not a bad thing. The new features in Windows have been rather user hostile and regressions in term of usability and capabilities. I’d be happy with them maintaining the security and adding features as required by advances in hardware, but not treating the OS as a playground for radical creativity and monetisation.
Right. It makes sense because more and more over time, the user's "apps" are in the browser. The OS is arguably more the browser and less the actual OS.
My wife, for example, plays games on Facebook, uses Gmail, and surfs a few websites she likes. That's why Chromebooks are so popular.
Windows is competing perhaps more with the web (and soon, WASM) than say MacOS.
Agreed completely. There's still a large group of people for which 2000/XP was when the GUI was already nearly perfect, and everything after that was downhill. As an OS I think Windows doesn't need any "new features" in the way of radical GUI changes anymore, just continued bug fixes.
Ditto for a lot of other software; enough with the churn, stability is more important.
Strongly agree. I quickly went from installing software on whatever OS I happened to be using with Windows 7, to actively avoiding booting into Windows with Windows 10. Now, if a binary is available for Linux, I'm using that.
Unfortunately, some areas like embedded development are stuck in the past with support for modern platforms. For example, it is somewhat ironic that I cannot use most tooling for baremetal ARM devices on an ARM platform (tablet, chromebook, raspberry pi) because Eclipse 4+ is a common dependency and is not built for ARM. And before you say it - go try to build Eclipse from source yourself in a reasonable amount of time. Nightmares.
With open toolchains available to serve most of my needs from coding to compilation to EDA to graphic design, it's becoming less of an issue. And on x86/64 platforms, more and more places are starting to provide multiplatform support. But it is so flippin' annoying to have to deal with all the cruft and crap of Win10 when all I want is a clean environment for creating things. Just make the OS easy to use and get it out of the way! It's a product, not a service!
Slowing down is fine, but lowering the amount of development effort overall is not so great. Windows 10, at least outside the kernel, still needs a huge amount of bugfixing and polishing. Mere maintenance is tolerable but not good.
Kind of frustrating to hear that the largest OS by marketshare for desktop PCs (82.68%) is going to be even less supported than it is currently. Doesn't seem like thats going to be good news for security on the windows platform.
Err, that's not true. In fact, the only reason they can charge for that OS is because they can keep it secure going forwards. It's not that Windows is being dropped, its more like MS realized that there is a lot more potential to grow in other areas, without being tied up to Windows.
They have failed to keep it secure for as long as I can remember. They are still able to sell it despite that, because people own Windows-only applications.
I would have agreed if you were talking about Windows XP but Microsoft has improved their security standing a lot since then. I think its a little disingenuous to call them unsecure more than other systems in the industry. Also the security environment has changed dramatically where monitoring and alerting are becoming key components. I know for a fact that Microsoft is investing heavily in those areas.
Microsoft seeing profitability and growth from its cloud business, and investing resources accordingly, doesn't mean its "putting Windows on the back burner".
In its cloud and enterprise offerings MS has been adding support Linux and open source in general. That mentality shift away from "MS products are for Windows OS" is more what's going on than "MS doesn't care about Windows anymore"
I would really love to see them rewrite Windows on a unix or linux core the way Mac did with Mac Os X. Having the software industry unified on a unix standard would be amazing. It would free up so much development effort currently spent making things cross platform (not all of it, but a lot of it).
Microsoft's understanding of what an operating system is and how it facilitates a user's interaction with their computing hardware has been completely misguided over the last decade. So much so that I can't tell if this headline describes something good, in the sense of salutary neglect, or whether it's a euphemistic summation of what's been done since Windows 7 (taking a considered, and profoundly user-centric approach toward OS design has been eschewed and given short shrift in favor of animated sprites in the start menu, for example).
They should take it off the burner altogether for Azure services and just use Linux. I’m not saying they should not host eg exchange, sql server, share point, and all their other stuff for customers. But much of their lower level infra is built on Windows, and Windows is both not ideal for large scale services and lacks the awesome FOSS ecosystem around it. And that’s before you consider how painful and slow it is to develop for.
I'm just a lowly small business owner, so my opinion isn't worth much. But if this is true, I feel it is a huge mistake. It feels like number chasing at best. Sure, Azure and online services (such as o365) are very profitable for the company, but I feel this is largely because of Windows not in spite of it.
If Windows erodes share or loses it's edge with corporations I feel like Azure & the online services will slide with it. But I don't have access to the data Microsoft execs do and I could be completely wrong. Maybe Azure is mostly Linux servers and o365 isn't really that important to their overall strategy.
Actually it shows that platforms really, really ought to be separated into at least two entities: “stable foundation” and “playground for clueless UI designers” or some other higher-level beast.
Any platform (Windows or not) in recent memory that underwent fundamental UI changes would have been improved if we had separated the layers. Let us continue to benefit from intelligent foundation improvements without being forced to inherit other changes.
There are lots of ways to draw these lines but there needs to be at least one line, below which software updates should have an entirely different meaning.
66 comments
[ 2.4 ms ] story [ 40.4 ms ] threadThat could disappear like a console generation. That's a little disconcerting.
[1]: https://en.wikipedia.org/wiki/HTML_Application
I use uTorrent on my mac.
Of course, then we'll have to deal with whatever the next platform is, but that's just the reality of the industry.
I fear we will lose the Microsoft as the silent steward of backwards compatibility, DirectX support, getting manufacturers to make great hardware drivers, the list goes on. It's this part of Microsoft that Windows will lose, because it isn't customer facing.
I think the fallout is exactly the opposite of what you want. A de-emphasized Windows means that the customer experience stuff is pushed forward while the above suffers. The only bulwark against that is how deeply invested most enterprise companies are in the Windows universe, driving many of those behind-the-scenes fixes and changes.
> Of course, then we'll have to deal with whatever the next platform is, but that's just the reality of the industry.
Gaming on the PC was, in many was, a repudiation of that idea. We could have one continuous open system that ran everything from old games to new games. That's part of the point the emulation crowd wants to make as well. We don't have to live with platform churn.
But I'll be the first to admit Stallman was right. We collectively ceded this power to a private corporation, and we could lose much of what we had because we weren't good stewards.
EDIT: Spelling and grammar because I lack coffee.
The vast majority of gamers using Windows buy their games through Steam, Origin, or other stores that Microsoft has nothing to do with. Unless we see a massive market shift then I don't see this having any affect on the future of PC gaming.
As for the game mode stuff, it's easy to turn off if you don't want it so I don't see any issue with it.
I also think that most gamers (myself included) would be okay with PC gaming becoming more consolized.
Back in the 90's you had to mess with configuration files in both the OS and the game you were playing far more than you do today. In a sense, it's because PC games have become more consolized - they "just work" most of the time. People like that. I hope the future of all forms of computing "just work" better than they do now.
I will concede that if you want absolute freedom of customization, Linux is the best OS for you even if you don't plan on gaming at all. However, saying that Linux is the future of gaming is about as accurate as all of the claims that "20XX will be the year of the Linux desktop".
Gamers that I know disagree and very much dislike this trend.
1) coreutils + some other nix commands
2) iTerm2 equivalent(ConEmu/Cmder do not compare)
DISPLAY=:0.0 xfce4-terminal &
After that, it about 90% of typical Unix setup with all the control key bindings for copy-paste, etc.
I wish WSL 's bash environment can run windows programs/utilities like Cygnus.
https://docs.microsoft.com/en-us/windows/wsl/interop
I tried insider for a few months and found the frequent updates too much, have some stability issues and stop it. Maybe it is a good reason to to start it again.
A second point, WSL programs don't rely on an active conhost shell to keep running. You can kick them off and run them in the background just fine. On one of my machines, I have a full i3 session running with no conhost window in sight. Everything works seamlessly, I can talk to my editor running as a Windows app, cut & paste back and forth, see file change notifications, hook up my docker command with my docker for windows VM, and so on. It's nice enough that I miss it when on my Mac and get annoyed with some of the fussy application alternatives on the typical Linux desktop.
It's still ages behind *nix OSs.
And I don’t care about the *nix implementations because the first thing I do is start tmux and use that. Same as I only care for the Linux userland which is what WSL gives me.
1) show the Git branch in the prompt 2) display nice colours 3) open multiple tabs
Looking forward to being proved wrong :) Cmder works well for me but I'd rather have something core to Windows to use instead.
Colors can be customized, but changing them is a PITA so we made (ColorTool)[https://github.com/Microsoft/console/tree/master/tools/Color...] to make it a little easier
Auto updating the prompt isn't possible in CMD unfortunately. I do believe it's possible in powershell, but that's not my cup of tea. I do have a batch script that gets the branch and re-sets my prompt. You could create a git.bat that calls such a script on every git command, or like me, create `vf.bat` that I use to cd and update the prompt at once.
Also in this picture: cmd.exe running inside tmux in WSL. Coming soon to Insider's Builds near you
Powershell is miles ahead in every way I can think of.
Also the whole office 365 plan gets F500 companies already tied into Azure. It's sort of like the drug dealer's "first one's free" tactic. They will, for example, spend money on direct connect network for O365, which makes Azure immediately more attractive. And their Active Directory info is necessarily already federated to Azure.
Look at what they've released in the past two years... their AI stuff is second rate compared to Google, and seems like its just there to check boxes. Their serverless stuff beyond Lambda is overpriced, like Aurora Serverless and Fargate. They're dragging their feet on EKS; it was announced 5 months ago, we still don't have a release date, and the latest webcast (last week) they showed made it seem either really bad or really unfinished, like you have to involve CloudFormation to provision instances, manually manage VPC subnets, etc.
Its a world where they have a core set of "legacy" building blocks that are really good and well-tested (EC2, RDS, S3, SNS, SQS, a few others) but if you go beyond those you enter uncharted territory that mostly exists to seem hip and check boxes.
GCP, on the other hand, has legitimately powerful products at higher abstraction levels like GKE, Firebase, DialogFlow, and their AI stuff. I know less about Azure, but they seem to be doing well checking boxes and pushing toward more higher abstraction levels with things like AKS, service fabric, etc. Where's the actionable innovation at AWS?
Though, I get some of the egress upcharge in GCP. Their global network is worth some premium. Just not the current predatory pricing.
My wife, for example, plays games on Facebook, uses Gmail, and surfs a few websites she likes. That's why Chromebooks are so popular.
Windows is competing perhaps more with the web (and soon, WASM) than say MacOS.
Ditto for a lot of other software; enough with the churn, stability is more important.
Unfortunately, some areas like embedded development are stuck in the past with support for modern platforms. For example, it is somewhat ironic that I cannot use most tooling for baremetal ARM devices on an ARM platform (tablet, chromebook, raspberry pi) because Eclipse 4+ is a common dependency and is not built for ARM. And before you say it - go try to build Eclipse from source yourself in a reasonable amount of time. Nightmares.
With open toolchains available to serve most of my needs from coding to compilation to EDA to graphic design, it's becoming less of an issue. And on x86/64 platforms, more and more places are starting to provide multiplatform support. But it is so flippin' annoying to have to deal with all the cruft and crap of Win10 when all I want is a clean environment for creating things. Just make the OS easy to use and get it out of the way! It's a product, not a service!
Source: https://en.wikipedia.org/wiki/Usage_share_of_operating_syste...
Microsoft seeing profitability and growth from its cloud business, and investing resources accordingly, doesn't mean its "putting Windows on the back burner".
In its cloud and enterprise offerings MS has been adding support Linux and open source in general. That mentality shift away from "MS products are for Windows OS" is more what's going on than "MS doesn't care about Windows anymore"
Microsoft's understanding of what an operating system is and how it facilitates a user's interaction with their computing hardware has been completely misguided over the last decade. So much so that I can't tell if this headline describes something good, in the sense of salutary neglect, or whether it's a euphemistic summation of what's been done since Windows 7 (taking a considered, and profoundly user-centric approach toward OS design has been eschewed and given short shrift in favor of animated sprites in the start menu, for example).
If Windows erodes share or loses it's edge with corporations I feel like Azure & the online services will slide with it. But I don't have access to the data Microsoft execs do and I could be completely wrong. Maybe Azure is mostly Linux servers and o365 isn't really that important to their overall strategy.
Let’s see.
Any platform (Windows or not) in recent memory that underwent fundamental UI changes would have been improved if we had separated the layers. Let us continue to benefit from intelligent foundation improvements without being forced to inherit other changes.
There are lots of ways to draw these lines but there needs to be at least one line, below which software updates should have an entirely different meaning.