ReactOS releases 0.4.8 with experimental Vista/7/10 software compatibility | Hacker News Reader

10 software compatibility

reactos.org

617 pointsbratao312 comments

This has been one of my favorite software projects to watch, because it's such an insanely monumental undertaking and moves so damn slowly. There were people working on this two decades ago, putting in tiny fixes and little bits of code and whatnot, and they did it even knowing that basically nobody would see their work for decades.

Hell, I have never seen a ReactOS installation in the wild, and I'm the kind of person whose friends install Haiku, Nix and NetBSD as their daily drivers. And the second-hand stories I've heard about people who did install React was basically "I was bored, I put it on a drive, played around for 5 minutes, and wiped it."

But that doesn't discourage them. Meanwhile, work keeps going on in the background. So many untold man-hours of thankless work going into the project, and the vast majority of that "hard" work with no payoff for years.

And now it's finally getting close to actual Windows, after decades of work, and soon (well, years, but still) people will be using it everywhere as a replacement for Windows.

This is one of the most important software projects in history. My hat is off to the ReactOS developers, and congratulations on the latest release.

[edit] Here are the names of the 75 programmers that have worked on ReactOS: https://github.com/reactos/reactos/blob/master/CREDITS

> soon (well, years, but still) people will be using it everywhere as a replacement for Windows.

No they won't. It's a fringe OS that requires technical expertise to install. Not only that but it looks like "old Windows". Users won't want to work with a Windows from a decade ago. Not only that but it'll have bugs and no support.

You'll see tiny pockets of people using it but as Microsoft gets closer and closer to simply having a free OS it doesn't seem likely anyone would ever use this OS for anything beyond "hey, that's cool!".

> Users won't want to work with a Windows from a decade ago. Not only that but it'll have bugs and no support.

Depending on the age group, I guess. I know plenty of people who would say that Windows had its peak usability around Windows 2000/2003 and only got heavier and more confusing later on.

If I know that ReactOS had a good chance of running most regular Windows applications, I'd install it on my mother's laptop in a heartbeat, and she would love to see the familiar, lightweight OS.

Exactly. In windows 2000 the UI was pretty much unified, a button looked like a button, window borders were consistent, the start menu had a simple tree structure, icons were ugly but they were iconic (they at least tried to represent something, noone thought that one day a grey letter will be an icon, or a user will look lie a circle and a line). No touch nonsense, everything worked with a mouse or a keyboard (mnemonics were visible by default).

Plus no Cortana bs.

I switched to Mac in 2002. Windows 2000/XP ’Classic’ is how windows should look to me. Whenever I come across the more recent releases I find myself totally handicapped.

I think it's been downhill from Windows 2000, XP was OK once you turned off the tellytubbies UI.

Frankly I can't think of anything they've added since 2000 to windows (except better security) that was an improvement.

I occasionally have to use Windows for work and I really hate Windows 10, it's just a chore to deal with when you only boot it every couple of weeks just on the updates.

Windows 2000 was a revelation when we installed it on an 30 screen network in a library/learning centre in place of Windows 98 (a long time ago, somewhere in England).

I think the 'seniors' amongst us may well welcome ReactOS together with those in small companies who run legacy industrial hardware from Windows PCs (depending on stability &c).

> * I know plenty of people who would say that Windows had its peak usability around Windows 2000/2003 and only got heavier and more confusing later on.*

Given Microsoft's mostly successful effort to make Windows more friendly to the user, I find it hard to believe any of these people have extensive experience with Windows 7+

Oh, but they did. There was a series of problems with Vista that pissed people off, then there was the quite-ok Windows 7, and then Microsoft screwed everything up with Windows 8. The irony is, they've managed to fix a lot of Windows 8 UX problems in Windows 10, but then the telemetry stories and in-your-face pushing of the upgrade from 8 to 10 left some people I know stranded at 8 and afraid to try out 10. I guess this'll change in the next hardware upgrade cycle...

I've been windows power user for many years before I switched to Linux. I can say that WinXP was one of the best windows and Win7 was okay and usable. Win8 and Win10 complete shit. Everyone I know had issues with it and had hard times using it. Asking if it was possible to use at least Win7 again. Windows completely screwed up there UI thinking it would be easier for people to use and in process completely killing windows. Not to mantion Win8/10 security/privacy horror storries. The last usable windows was Win7. This is why I see many people switching to Mac OS X or Linux.

> WinXP was one of the best windows

On a somewhat more objective note, there was a lot of criticism back when it came out. (Specifically, about "the bloat". Everything is relative, I guess; one can say that things continued to get worse and worse from Windows 2000 on, although some would swear that it was NT4 that was the best Windows ever - e.g. it ran in 16MB of RAM vs. Windows 2000 that raised the requirement to 32MB; for comparison, Windows 95 only needed 8MB.)

Vista and 8 had their problems, but 10 has been fine for me. I disagree with the statement "The last usable windows was Win7." But then, I've seen the same argument before, in different flavors. People complained about Windows 7, wondering why anyone would switch from XP. They complained about XP, wondering why anyone would switch from 98SE.

I feel like "quite-ok Windows 7" is underselling it. Response was largely positive and it wasn't even that different from Vista. It simply didn't suffer from the same problems, since hardware manufacturers caught up.

Yeah, Vista had two big issues that had both been mostly resolved by the time 7 was released: 1) Driver support was poor, and 2) A bunch of low end computers with "Vista Capable" stickers that really meant "It should boot but it's going to be awful."

IIRC Microsoft was pressured by OEMs into making a lower tier of Vista qualified hardware even though they knew it really required more than that to work acceptably well.

I never personally owned or used a Vista machine, but my office has a couple of them plugged in for a showroom area where they cycle powerpoint slides, and oh my god is it painful if you ever have to touch them. IIRC they've got 512 MB of RAM, and some of that is being shared with the iGPU. It's not pretty.

I used two sets of Vista machine, the 32-bit machines were almost all terribly slow, the 64-bit machines were surprisingly capable. So much so, that I could use a Vista 64-bit machine as an alternative to Windows 7 if it were still supported. The machines sold as Vista-capable weren't all that great with XP, but better, so were even under-powered for the previous generation. Vista leaned a lot more heavily on disk I/O that its predecessors, so turning off Windows Search, and a few other things I forget made it tolerable to use. It must have been about that time that all of Microsoft's developers got SSDs on their desk (I'm only half joking).

I remember installing Windows 95, from floppy disks back in 1996 or so. The system requirements on the box said 4MB of RAM. And it was true, Windows would run with 4MB. In the fine print, the box also said if you wanted to do anything besides play solitaire, like word processing, 8 was recommended.

So it's been going on a long time, the lowball estimate of systems requirements to get you to buy the product, then the real cost after you've already installed it. 4MB is a rounding error these days, back then it was at least a couple hundred dollars.

Vista being painfully slow was not limited to low-end machines. I had a fairly powerful (8GB+ RAM, new i5, etc.) gaming rig when it came out, and everything was slow on it: I'd click something in the interface and have to wait 5+ seconds for it to respond, Windows Explorer would crash randomly, etc. Even booting up seemed to take forever. Upgrading to Windows 7 fixed all such issues, and I still have the machine to this day, which I've kept with Windows 7 (although I upgraded the GPU).

Damn.

512MB wasn't even enough for my WinXP box because the softeware apart from the OS became more demanding with time.

They fixed some UX problems but left others. Two control panels, massive grey buttons that feel odd when used with a mouse, dysfunctional and random start menu, full screen modal notifications (windows server 2016), etc

The two control panels thing is maddening. It's just embarrassing they've let this mess escape beta. It should be one or the other, not an unevenly applied mixture. The lack of cohesiveness makes the whole system look amateurish.

I am an heavy keyboard user, and prefer it to the mouse.

I find that Windows 2000 has the best support for keyboard users among GUIs. Better that MacOS, better than all the incarnations of Gnome/KDE, better that the following releases of Windows, that "forgot" about the keyboard and focused on other input systems.

You can see that in many small things, like sane keybindings and reasonable sequence for selection through tabs (for example, going from the path bar on the top to the folder content requires two tabs in Windows 2000 but four in Windows 7, unless one does some find tuninng of the UI).

KDE3 > Windows 2000. It was truly amazing. Ripping CD's by copying folder. FULL network integration on the file manager. KIOslaves. Artsd sucked, ok.

Things that you mentioned are still fully functional btw.

I'm one of them and have as much extensive experience with Windows 7+ as someone who write *nix and embedded software for a living can have. Windows 2000 is the last Windows version I used exclusively.

The "mostly successful efforts" you speak of seem to have been focused on delivering ads, TOS-backed spyware and those weird Metro apps. The result is certainly simpler from a strictly visual standpoint, but more usable is something that I wouldn't be too willing to concede without some data (and I don't mean install data/user base, given how Microsoft's customers don't have much of a choice in this department). Frankly, I don't see how the weird start menu or having both the Control Panel and some weird Settings thing help with usability, but UX is a surprising field, I guess.

That's a compromise. Settings is deep-searchable and unified between platforms. Control Panel is still there for backwards compatability (including with MS's own stuff)

That's bloody terrible, whenever I need something I have no idea where I'll find it.

I usually click the start button on my keyboard and type what I'm looking for... most of the time, it's the first option that comes up. It mostly works for apps too, excluding some annoying MS app placement issues. I've moved on... Windows 10 was the last straw. Not the telemetry, but the stupid integrated ads, especially the ones over the chrome icon in the launch bar.

Windows 8 was absolutely dire. Windows 10 is very good, but not in any way brilliant.

> It's a fringe OS that requires technical expertise to install.

That's fixable (and not different from Windows).

> Not only that but it looks like "old Windows"

That is by far the easiest fix of all. People just care more about making things work right now.

Plus, a lot of people actually preferred the older style.

> Not only that but it'll have bugs

Welcome to software. Windows is also full of bugs.

> You'll see tiny pockets of people using it but as Microsoft gets closer and closer to simply having a free OS it doesn't seem likely anyone would ever use this OS for anything beyond "hey, that's cool!".

I think it might end up being quite big quite soon: There's a bunch of proprietary software out there that is old, and not Windows 10 compatible.

Things like car diagnostics software. If ReactOS gives them a way to get things working without Win10, I'm sure they'd eat it right up.

ReactOS just isn't ready for that yet. It's still major version 0.

I know a bunch of people (mostly gamers) who refuse to upgrade from Windows 7 to 8 or 10, mostly because of usability and privacy concerns.

Eventually, they'll have to (DirectX 12, but I'm sure there's a ton of APIs for non-games as well and drivers will stop being updated for Win7).

But if ReactOS can reach a level of maturity and compatibility to please such people, there's a sizeable niche there of people who:

  1. won't switch away from Windows to Linux or BSD,

  2. but would be willing to adopt a familiar and compatible operating system.

Please don't use code blocks for normal text. This is frustratingly difficult to read on a phone where the code block becomes a scrolling container.

Sorry, I didn't intend to make it a code block, just wanted to indent the list headers. Is there a list syntax for HN comments?

Not really, but it works well enough if you put each item on a separate paragraph (i.e. with blank lines before and afterwards) and prefix it with a bullet point etc.:

- Then it looks like this.

- And still flows nicely.

I definitely count myself as one of those gamers who hasn't moved off 7 yet for my gaming and music production. On a machine like that, windows is just a layer between the game and the hardware and the less other stuff, the better. 8 and 10 have little value add for someone who just wants program compatibility-7 was plenty stable.

8 and 8.1 had huge performance enhancements, though. Easily the fastest Windows to-date.

To an average user, the latest Windows versions really don't add anything useful over WinXP or even Win2k. Many things got worse, like discovering and starting installed applications. Slap a modern UI theme on XP, and nobody will notice that it's a nearly 20 year old OS.

I have high hopes for ReactOS, because pro audio software is currently the only thing that really keeps me tied to Windows. (I use GNU/Linux for work already for more than a decade.)

Unfortunately, last time I looked audio software and drivers weren't well supported on ReactOS. If I could get Reaper, my audio interface and commercial VSTs running on it, I'd ditch Windows forever.

Not to invalidate all of your arguments, but it does “just install” in a VM without needing technical experience.

To reach any mass of users (outside of turnkey solutions where the end users may not know or care), you need commercial support and the later needs concrete identified market segments. To parallel Wine - ROS needs its own CrossOver (but not Cedega).

I agree - I literally downloaded the ISO, booted it up, it installed in VirtualBox.

Of course, the network driver didn't work, sadly :(

change to use pc net iii, there's a driver for that built in

> Not only that but it looks like "old Windows".

Not convinced that's a weakness as opposed to a strength. Every time Microsoft changes literally anything about their UI, they alienate a legion of people who suddenly feel their workflow and productivity has been irretrievably broken.

I currently use win 7 in win 98 style. It is faster on the graphics and the processor.

I can see ruining old but required Windows binaries without paying for a Windows license as a totally valid use case.

I installed ReactOS in a VM, because I have one Windows app I need (Adobe Digital Editions for reading library ebooks), and I figured it would be easier than trying to get it running in Wine. But the current ADE requires Windows 7, so I was out of luck. Probably worth trying the new version.

People are more likely to use Linux with Wine to run some Windows programs. Since it actually produces a working result.

But if ReactOS can be practically comparable - it can get some traction too.

IIRC Wine and ReactOS are actually closely aligned. Many improvements made in ReactOS are ported to Wine and vice-versa.

I think you're entirely right, but I wonder if there's not some life left in it down the road. Whenever I've needed DOS, which isn't that often anymore, I reach for FreeDOS, not MS-DOS. Who knows what deeply legacy systems might be around in 30 years? But then again, it's hard to imagine something as ubiquitous as a Windows installation disk being unobtainable on the used market.

> as Microsoft gets closer and closer to simply having a free OS it doesn't seem likely anyone would ever use this OS for anything beyond "hey, that's cool!".

Free OS? Even bog standard windows 10 pro (non-oem) is hardly free. Never mind a datacenter server license (aka you need to run 10s of instances in vms in order to get anything line useful utilisation from this modern server).

Windows have become bloated to the point I have rather installed openSUSE and after ~20 years I am not touching them any more. I don't like linux (it has another set of problems) and I would be very glad if I could use ReactOS and instead using firewall, just recompile the winsock dll with my own instructions (just one example).

Windows will never be free os, you will pay heavly with your privacy and from what I have seen on GUI changes, they are more moving into direction of cloud (minimal installation on pc, the PAID resources on cloud) than beeing free.

ReactOS devs, good work and thank you, I hope I will be able to use it in the future.

> Windows have become bloated

I don't agree. Vista was a mess but they seem to have made such tight performance budgets for 7, 8 and 10 that any computer which can run the 11-year-old Vista can also run 10.

10 is bloated by most metrics. Applications you don't need or want are bundled by default, it has a 20+GB install size, two separate sets of overlapping configuration utilities, a built-in HTTP server, built-in adware, etc.

There has been a lot of good work done on the underlying system, but it's had a ton of useless buggy crap piled on top of it.

Depends on which definition of bloated you choose. If you look at install size on disk it's clear that Windows 10 requires more space than Windows 7. In contrast, you can easily create a usable Linux system (including a window manager, not just a minimal server install) that takes up half the space of a Windows 7 install.

I do not need Cortana, XBox, Store, Telemetry, One Drive ... Can I remove them? Microsoft says no. Microsoft has become the definition of bloatware.

Just check kernel32.dll, it was CLEAN for years, now there are references to .net framework and it is becoming a mess. Believe me, Windows are becoming a mess, I never liked linux (actually I hated it) but the directions that Windows are taking are worrying (it is the same perspective when I say linux is becoming a bloated piece of , with each release it is becoming worse but at least it is open source. But dont get me wrong, I am writting this from purely technical perspective, for a typical user, everything is just fine. No need to worry, except maybe Windows 365 in 2020.

(I will refrain from further commenting as my technical perspective is hitting into large fanbase of both systems and karma here is really not beeing able to handle it. At the end it is the terror of the less technical average.)

The dependencies with long DLL names beginning with api-ms-win- are unrelated to .NET framework.

Here’s more info: https://scissortools.wordpress.com/2011/11/23/more-on-window...

I am not talking about dependencies, I am talking about COM exports from kernel32, than manifests, multiple versions of same library on the system which now matches linux xzy.so,xzy.so.1,... they had a clean userspace with backward compatibility, now they are having a mess.

Two things

>It lights a fire under M$'s asses to make windows competitive

>Adoption is slow but exponential. Today is the first day I learned about ReactOS and I'm excited. I can program and I have money. If people can find value in it, adoption will grow.

>It lights a fire under M$'s asses to make windows competitive

Err... no. Not unless the installed base reaches a critical mass. And certainly not when the market share is not even 0.1%. macOS (or even desktop Linux) has a higher installed base and they don’t bother Microsoft significantly.

I’d be more optimistic if the OS had not been around for two decades. It had its chance when the likes of Vista was released. If it didn’t make a dent back then, it is not going to make one now - especially since Windows is at its best in terms of usability. Sure there are concerns about privacy, but those didn’t prevent Google from acquiring a (almost) monopoly in mobile market. Microsoft’s own attempts went nowhere — and they have infinitely deeper pockets compared to people behind ReactOS.

And even if it does gain momentum, I am fairly sure that Microsoft will find a a way to sue them into oblivion.

> (or even desktop Linux) has a higher installed base and they don’t bother Microsoft significantly.

Except for they did. I'm pretty sure WSL was about developers who had preferred GNU/Linux-based systems (or macOS, since it can be compatible for some cases) for work, not about just running some *nix server software on Windows machines.

I said “significantly” — macOS and Linux have existed for ages, yet Microsoft is starting with WSL just now. Also, as far as I know, WSL is a by-product of the Poject Astoria - not something Microsoft set out to do. They probably thought, “We’ve done most of the work anyway, might as well get something out of it.”

So it will become this is the year of ReactOS on the desktop?

Windows 10 is good (really!) if you haven't taken a look at in a while. I switched from Mac on the Desktop to Windows 10 because I needed "pro" hardware (NVIDIA GPU, 1 TB RAM, etc) and it works just fine. They've really been on the ball in the last couple of years.

It’d be good if it didn’t have all the damned spam baked in with non obvious ways to remove it. I got so mad the FIFTH time windows 10 pro installed Bubble Witch Saga and Candy Crush on my system.

After seeing that I really want to save up and buy only Macs from now on.

Why not just run a FOSS OS?

A fair question, I actually did install two different Linux distributions a couple weeks ago. For some background I'm a software developer, and consider myself reasonably competent in using Linux.

I started with Fedora. Installed, seemed nice. Then I did a dnf update or something like that. It froze for 4 hours in the middle of the install. So I hit ctrl+z. When I restarted my computer GRUB was attempting to boot an OS version that did not exist and I had to manually try to figure out how to fix this which I was in no mood to do after having the OS for 10 minutes. Also apparently the newest version of Fedora didn't have the concept of turning off auto lock? Or a timer. I'm not sure, but the end result was that if I tried to watch a movie for more than 10 minutes my desktop would helpfully lock, continuing to play the movie.

So I then installed Ubuntu, which was fairly straightforward. Then deleted the couple of things that felt like advertisements (like a direct link to Amazon on the quickbar or whatever it is). Overall, was a pretty good experience until I installed the proprietary Nvidia drivers. Then for some reasons movies (DRM-free files) would stutter. At that point I was thinking about doing some C# development in visual studio anyway so I looked into dual boot installing windows 10 while having Ubuntu installed, and while there are a million articles on how to install Ubuntu with Windows 10 already installed, the reverse did not seem to be true. So I just wiped my system drive and reinstalled Windows 10, which helpfully activated based on my hardware profile even though I'd neglected to save the key.

Installing Windows first is just the recommended way to set up dual-boot, since the Windows installer will always overwrite the bootloader without asking any questions. I think these days there's an automated repair tool on the Ubuntu install media that'll take care of fixing it, though. And with Nvidia drivers it's usually a good idea to install the latest ones from the "graphics-drivers" repository (there should be plenty of guides for that too).

Also, depending on what kind of C# development you do, VS Code can be very nice. DotNET Core is very straightforward to work with, and targeting the full frameworks with Mono is doable with a little bit of tweaking. If it's Unity, then the debugger plugin will be completely useless on Linux, and you'll have to use the very latest beta version of the editor.

If you ever feel like trying Linux again, perhaps try Manjaro? It's a good place to start for a new user, and more polished and stable than Ubuntu and Fedora by far.

Hardware support of laptops for FOSS is not that good unless you restrict yourself to Dell or IBM.

It hasn't been the case since a long time really. I have used it on enterprise HP and Lenovo laptops and everything works flawlessly.

A strange argument: hardware support of laptops for MacOS is officially non-existent unless you restrict yourself to Apple. As soon as you're willing to consider 'Hackintosh' installations you should also consider Linux distributions as those are generally far easier to install and maintain.

While I would say that throwing linux on any random laptop and expecting perfect compatibility is foolhardy, it's really not that difficult to find a linux compatible laptop these days.

In general if you stay away from anything too esoteric, and avoid hardware that needs proprietary drivers (e.g. nvidia graphics, broadcom wifi etc...) most things should work out of the box, but it's a good idea to do a bit of research on anything you're thinking of buying if linux compatibility is desired.

There are also various smaller vendors like system76 that sell laptops with linux pre-installed.

Also has very good support on Lenovo (in my experience), which would make 3 of the largest PC/Laptop manufacturers in the world that has good linux support.

Here is a list of certified vendors for Ubuntu: https://certification.ubuntu.com/desktop/

IBM no longer makes desktop PC s or laptops. Sold the designs to Lenovo years ago.

I had HP and Compaq laptops that ran Linux as well.

That was true 5-10 years ago, these days it's much better (except for some old offenders like Nvidia GPUs or Broadcom WiFi modules). My Asus notebook never lacked any support even though I bought it just three days after that particular model was first released.

My HP laptops have also had zero problems running Linux for years.

Not OP, but for me the only hurdle left is gaming compatibility.

> Windows 10 is good

Did Linux succeed because other UNIXes were bad?

In my opinion, the reason why we need ReactOS is independent from whether Windows 10 is good or not.

And yeah, I agree besides the privacy issues, Win 10 is pretty good

> Did Linux succeed because other UNIXes were bad?

No, it succeeded because there was a need for a UNIX like OS not encumbered by lawsuits (like BSDs) or costing a fortune (Solaris, SGI etc.).

Linux is a clean room implementation that doesn’t borrow any code from proprietary AT&T UNIX, ensuring that it has future. That, and the backing of FSF means it is not going to be sued into oblivion - something that could not be said for BSDs and other free variants back in 1990s. No one wants to invest into a something that may be shut down anytime. It is the fighting thtpat hurt the BSDs the most — causing them to be overshadowed by Linux despite being superior in many ways back in the days. Some would argue that they are still technologically superior in certain areas — but that is debatable.

Linux succeeded because bsd had legal issues.

People like to bring that up, but Linux also succeeded because of an open development model that BSD lacked. Linus Tolvards is basically the poster boy of Open Source.

For desktop, I use opensuse (just clarification, to avoid holly wars here).

I have migrated my servers to FreeBSD not due to problems with kernel, but due to everything else, the distributions are really doing horrible job maintaining userspace, not to mention failures like systemd. Also trusting data to linux is an interesting ride, from btrfs failures to "merging" (pun intended) those failures to zfs (https://github.com/zfsonlinux/zfs/issues/7401). FreeBSD has much slower pace of development and this is good for stability. Zfs is rock solid beeing part of OS for decade, docker was ported to it in 14 days, due to jails which are there for 15 years+ and field tested for years (while I was listening from linux guys years that chroot is all you need). I wont complain about linux features, but for BSD, I know that after update everything will work exactly like it worked before and this is more important for my bussiness than all the bleading edge features.

And then you have all the other issues that Linus explains here, for debian but same problems other distributions have (and yeah, I am just a bit younger and highly proficient in c/c++ on multiple operating systems, developing system level application for more then 20 years, that is probably why I share his opinion): https://youtu.be/1Mg5_gxNXTo?t=459

Linux has succeeded as it was fast by cutting corners in development. And it worked for a while, now problems started to stockpile and turtle is starting to catch up.

Regarding the "open model", I don't understand what would that be, BSD license is more permissive than GPL? Not to mention a shame when (https://www.zdnet.com/article/linux-beats-internal-legal-thr...) kernel developer for linux started suing companies not releasing the source code.

Just to wrap it up, if ReactOS will be able to keep the pace of development with the Windows changes it has a good chance that the year of ReactOS on desktop will be much sooner than the year of Linux on desktop. And this is what we want, not Linux OS on desktop but Open source OS on desktop, right?

Yes, Linux did succeed by being better than other Unixes in ways that count on the server side.

And on the desktop it took a backseat to MacOS for similar reasons.

Have you tried installing MacOS on non-apple hardware? Not sure I understand the comparison.

MacOS shines as long as you are within Apple walled garden, while I personally like MacOS, I know plenty of users who like Apple hardware but not the OS.

Windows or Most Linux distro doesn't have any such limitation and will work out of the box 90%+ of the time.

I am not going to dispute your claims, MacOS does not run on non-Apple hardware.

What I am saying is that many people that would have been Linux users are now MacOS users. And one of those people is me.

I was a hardcore Ubuntu / Debian fan, but then got a MacBook Pro from the company I worked for in 2014. After moaning about it for a while, I'm now a convert. Initially it was because of the hardware, as I love Retina displays for example.

My younger self wouldn't believe what I'm about to say, but things break a lot on the Linux desktop and I don't have the patience for that kind of crap anymore. Not sure how it is nowadays, but back when I was using it, the presence of an Nvidia graphics card meant any upgrade could break your setup. Which isn't acceptable given Nvidia's popularity.

Many times it isn't Linux's fault of course. Surely you can't blame Linux for Skype not working, but the general availability and quality of apps is a big problem. Also I agree that you can't expect Linux to work flawlessly on all laptops and some research is needed for a good experience. Thinkpads are in general good and some Dell models too. But this information is scattered all over the place and translates in even more time lost.

Go to any conference and you'll see a majority of people with MacBooks. Those people could have been Linux users instead. They were Linux users only a few years back.

I'm not saying that I'll never use Linux anymore. I keep using it on the server-side and I might be back to a Linux laptop someday, due to Apple's newfound hostility to power users.

But just an FYI, the freedom argument doesn't really win me over. If all other things are equal, I prefer the open source solution of course, but I've got work to do, bills to pay, projects to finish and a family to take care of, so until Linux improves to the point where I can do those things without wasting time, I'll stay on MacOS.

..once you fix a few things [1]:

> This Windows 10 Setup Script turns off a bunch of unnecessary Windows 10 telemetery, bloatware, & privacy things. Not guaranteed to catch everything. Review and tweak before running. Reboot after running. Scripts for reversing are included and commented.

[1] https://gist.github.com/alirobe/7f3b34ad89a159e6daa1

Very guaranteed to not turn off Cortana in fact. Goodbye 300 MB of RAM.

I use Fedora on the desktop but keep a copy of Windows 10 in VirtualBox for times when I need to use Photoshop (I'm a web developer) or some other Windows-only tool.

I recently updated to the "Creators Update" and couldn't believe the amount of shit that popped up after the install asking me pointless questions, trying to integrate with things, trying to get me to activate OneDrive etc.

It's not just post-install either, this stuff pops up when you're in the middle of something else!

The thing I like most about Gnome is how it just stays out of the way. I don't use my OS, I use the applications I have installed. The shell should provide an optimal environment for them to run in and an easy way to launch them - and nothing else.

Don't get me wrong, I've been a fan of Windows for a long time and 10 is the best version for years, but the popular Linux distros and macOS seem to do a better job of staying out of the way of the user.

I think Microsoft are losing track with all the "value added" stuff they're constantly trying to foist on people - and don't even get me started on the tracking.

It is good. The privacy concerns are real though.

> Hell, I have never seen a ReactOS installation in the wild

It's not especially common, but I do know of a couple kiosks and a PoS system that run ReactOS under the hood to avoid Windows licensing.

Russia is using it.

https://www.reactos.org/project-news/reactos-second-os-russi...

How the hell do they get PCI certified with ReactOS?

I don't have details, I just happened to see a crash to desktop, but it may not have been certified - many Australian retailers will pay the fines rather than go through the effort of ensuring compliance.

I think it would be pretty rare for Australian POS machines to handle credit card details. Almost everywhere I only see these systems using a separate terminal with its own network connection to handle the card transaction. Usually there is integration (the POS system tells the terminal the price and to prompt for a card) but I don't think the card data ever touches the POS terminal itself (unlike the way they seem to work in the US).

This is a pretty good system, because the terminals are able to be updated really easily. A lot of cafes and stuff also use iPad POS software (one called 'Vend' is really popular here) and it doesn't have to be certified. This is part of the reason that it took hardly any time for almost everywhere to support contactless six or seven years ago. For example at Myer the POS systems look 15 years old but the attached card terminals are usually only a year or so old.

The iPAD solution is certified by the vendor; these employ a PCI P2PE certified terminal and a PA-DSS certified PoS software on the device. The software enforces security controls on the device as well as performs checks such as root/jailbreak detection, iOS patch level, security/passcode settings etc. and if any of these do not match what the vendor specifies (which is what they certified) it won’t work.

As for the terminals on older PoS as long as the PED is certified and the PoS is certitied its not a problem.

Ah the Tesco method. :)

Typically credit card processing is handled entirely by a physical device plugged into the computer. The computer isn't in a PCI zone.

The POS is very much in the PCI zone, the PED and card readers will be certified separately.

If the POS doesn't touch card data how would it fall under PCI?

Card data isn’t the only data that is covered by PCI SSC standards.

Card holder PII is also covered and is even considered more important these days since CC numbers are easy to rotate but your identify isn’t.

Also even if the PoS doesn’t sees the card details it is part of the payment acceptance process and if it’s compromised the payment process can be affected even with P2PE devices.

If the PED is complete separated from the payment process e.g. those in which the vendor has to type in the amount separately and the PoS doesn’t take any any any customer PII ever you may be able to get away with using something like ReactOS on it.

If the pos is system is regarded similarly as a cc accepting website that proxies cc data to an endpoint, then the os shouldn't be a variable of pci compliance

Most (European) terminals don't even proxy to computer, they're completely independent devices connected to wifi that communicate directly with bank. The connection to computer is used only for "1 EUR" and "OK"/"FAIL" kind of messages and are completely optional.

Just curious, can you share a few names?

Only worked with the kiosks directly, which were being deployed by Gumtree-Kiosk to medical clinics, and NEOTouch to retailers.

Both very small Australian companies, probably about 100k deployments between them.

> Hell, I have never seen a ReactOS installation in the wild

I've seen it once so far, because my employer has a no Windows policy for contractual / security reasons and I was requested to translate a foreign-language dialog (on an English language OS installation). It was to solve / use an esoteric bit of Windows, that ultimately did not succeed. I commend the effort to make such a distro, but it is still a decade or longer away from the "casual" techie.

Why would one use ReactOS over Wine? I read this page: https://www.reactos.org/wiki/WINE, but it didn't really provide an answer.

I understand that ReactOS doesn't require Linux as a dependency, but is that a significant win?

Wine and ReactOS are practically sister-projects at this point that share a ton of code.

ReactOS does replicate the NT Kernel, so in theory (if it ever gets there), ReactOS would be able to run Windows Drivers. IE: the "real" NVidia and AMD Drivers or any other kernel-level drivers that exist in Windows land.

These days, AMD (and even NVidia) have decent Linux drivers. And WiFi drivers have Windows->Linux translators as well. But full compatibility to any Windows binary (including device drivers) is certainly a noble goal. There are a lot of little hardware devices that are written for Windows only (ie: specialized medical equipment, CAM / CNC Mill programs, etc. etc.) ReactOS would allow a smooth transition to Open Source if they achieve their goals.

------------ EDIT: It should be noted that there are a ton of Windows OS-level details that Microsoft does much better than Linux too. IE: I'd argue that Microsoft's security model (SIDs + ACLs) is superior to Unix-style Users + Groups. A large group of open-source developers who strongly understand the low-level internals of Windows is certainly a good thing in any case.

You know you can use ACLs on Linux kernels and filesystems?

https://wiki.archlinux.org/index.php/Access_Control_Lists

Yeah. But Windows NT's ACL is kernel level and applies to any object that the NT kernel can make. (aka: anything with a "Handle" can be controlled with Window's security core). MMaps (called "Sections" in Windows), Threads, Processes, Mutexes, and more are covered by Window's ACL security model.

Even then, ACLs are just part of the picture. Windows NT's "SID" system for identifying user permissions is far more flexible than users / groups.

WinVista and later also adds mandatory access control on top of that, in particular event logging / auditing guaranteed by the kernel and "integrity levels". Anything that is "drive by downloaded" by Chrome for example has an integrity level of "untrusted" and thus is locked out of all kernel objects. (IE: "untrusted" integrity in Windows prevents access to files, processes, threads, mmaps, all services, etc. etc.).

----------

Linux's security model is too weak to be used in a modern operating system. That's why Google extended it with all of the Android App stuff. Android's security model is closer to what modern Windows can do.

And if you know anything about Android (despite being built on top of Linux), its security model is quite different. Windows basically offers Android-level security at the kernel level. (maybe a bit more: major services on Windows, such as LSASS aka Login / Password service, can run in an isolated VM for example)

Yes but it certainly isn't as flexible and for networking computers requires ensuring uids and gids don't clash with local users. There's talk of replacing posix acls with nfs4 acls, which were modelled on Windows ACLs.

Linux has ACL too.

Traditional Posix ACLs have quite different semantics than NT's ACLs (Even if you don't grant "Bypass traverse checking" to Everyone).

The Linux equivalent to NT ACLs are called RichACLs, and they are quite new.

ReactOS is meant to be a drop-in solution. Just install it over your existing Windows installation and keep using your machine like nothing has changed. That's the goal anyway.

Yes, if it installs correctly* on bare metal then one can boot to an explorer-like shell in a matter of seconds without going through a Linux desktop.

But I can see an obvious use case - testing. Customer reports an error but you can't reproduce it on your company's Windows development machines. I was able to reproduce a couple of bugs that way with wine and I imagine it could be handy running under ReactOS.

* I've only ever tried it in a VM.

I like what Linux is trying to do, but I havent been satisfied with desktop usage.

Ubuntu server with LAMP is the greatest thing of all time, but desktop Linux has disappointed me over the nuances like mouse acceleration settings, netflix not working natively, etc...

If you have problems with those details on Linux, I don't think you will be a fan of ReactOS.

Since it barely works at all.

Not sure what you mean with "Netflix not working natively" but on Linux you just need either Chromium or a recent version of Firefox (maybe not the ERS version) and enable it to play DRM content. It just works.

That hasn't always been the case, perhaphs he tried it a few years ago when you had to have silverlight.

I watch often TV on web sites on Ubuntu. Almost all issues have been solved in 2017. Before, it was failing more often than succeding.

For what it's worth, today, Netflix works perfectly, at least in (recent versions of) Firefox. The first time you play a video it asks to enable DRM, you click "Accept" and the video plays.

I'm not sure what issues you've had with your mouse in the past, but in the vast majority of cases it works out of the box. I can say that desktop Linux has come a VERY long way in the last 3-5 years.

Wine is in practice more useful. You can run demanding Windows games with it on Linux. So its wider usage is expected.

How long did it take for Windows 95 to be developed? It would be an interesting contrasts between Enterprise and Charitable development.

It's much easier to Greenfield your way rather than reverse engineer (bug-for-bug) complicated code.

Windows 95 still borrowed a lot of code from MS-DOS, though. That's why you could still exit to the command prompt. I think you could do that through WinME.

Huh? Why would you even say this? There's no MS-DOS code in Windows 95, and you couldn't exit Windows 95 to a command prompt. Are you just trying to throw some FUD into this conversation?

No, he's right. Windows 95/98/Me use MS-DOS as a bootloader, and Windows 95+98 still allowed you to exit Windows from the shutdown prompt, leaving you only at the DOS prompt. Windows Me took away the ability to exit, least without a binary hack to IO.SYS that can re-enable it.

You mean DOS Mode where Win9X GUI shuts down and reboots into DOS mode to run DOS software. I was a Win95 beta tester.

The original "It's now safe to turn off your computer" screen in Windows 95/98 actually sent you to a DOS prompt. Because the computer was in graphics mode you couldn't see the prompt but you could blindly type the commands to switch modes and continue to use the computer from DOS.

I didn't remember it either, but Windows 95 actually allowed to exit to DOS.

https://guidebookgallery.org/pics/gui/startupshutdown/shutdo...

You could even configure Windows 95/98 to boot directly to DOS by setting BootGUI=0 in MSDOS.SYS, and then you'd have to run "WIN" to boot Windows

Isn't MSDOS.SYS a binary? Are you sure you don't mean CONFIG.SYS?

According to Wikipedia, it became a configuration file in Windows 95 (MS-DOS 7.0): https://en.wikipedia.org/wiki/MSDOS.SYS

Interesting, never knew that.

That would've made one of my favorite tricks impossible. Here's what I remember doing about those pesky login screens asking for passwords:

1. Reboot the computer in DOS mode.

2. Change to Windows directory.

3. Delete the .pwl file for a user.

4. Reboot back into Windows 95.

5. Enter a new password in what was once a login screen asking for the old one.

Simpler times for hackers back then. :)

when it starts getting really good, Microsoft will probably sue it to the ground.

How's that? It's open source and they wrote the whole thing from scratch. Even if they abandon the project it's open source and work can continue even after MS has their way with them.

Microsoft has more budget for making coffee for the legal department than ReactOS has budget. Even if the lawsuit has no chance of prevailing, it still can force the ReactOS team to shut down. Oracle made a convincing enough case against Google about copying APIs, IIRC.

You forget about the bigger FOSS community goodwill that they've carefully been trying to cultivate in the past few years that they will throw away by doing that.

I don't think they value goodwill over protecting their core business.

this is the reason why I wrote "probably" and not "will"

They'll try to claim to have patents on key functions, API copyright like Oracle, or look and feel with either type of law. IBM destroyed a company that used Hercules emulator to replace its mainframes with compatibility or competition arguments attempted with failure. Oracle vs Google on API issue shows what can happen when big money gets thrown at lawyers. For patents, Microsoft collecting over a billion dollars on royalties from Android vendors and Apple blocking Samsung in Germany over look-and-feel are examples.

Anything trying to be compatible with a large, greedy company's software is a huge risk if it starts cutting into their profits. Maybe nothing will happen but something might happen. I'm more concerned for companies like EnterpriseDB than FOSS projects like ReactOS, though. The lawyers do prioritize on those making money with the competing software.

Honestly, I don't think we'd try to destroy ReactOS at all. Windows has stopped being a cash cow for us - recently, we've even started to disband the Windows org, moving the kernel to Azure etc.

I wouldn't be surprised if we open-sourced NT, at the rate things are going. Windows devs are likely flattered by ReactOS at this point.

You have a good point. They might just try to suck a lot of the money out of the businesses built on ReactOS like they do with Android. Then again, Android wasn't going to eliminate a lot of revenue from locked-in customers like ReactOS could. I say them forcing royalties is most probable outcome for now.

They've had problems in the past, based on the legitimacy of reverse engineering Windows: https://en.wikipedia.org/wiki/ReactOS#Internal_audit

> Even if they abandon the project it's open source and work can continue even after MS has their way with them.

Such infringement claims are damaging to the legitimacy of the codebase itself. Anyone taking up the project could face the same problems.

In Russia? Yeah, right... :D

I'm thinking seriusly installing on an old computer that haves Windows XP and a Debian 9. I have Windows XP there, because is the last Windows that supports a old Roland plotter from early of 90's. ReactOS can use Windows XP drivers ?

I saw (back in 2012) a ReactOS deployed at points-of-sale in large retail chain in (then Ukrainian) Crimea. Maybe it was cheaper to make their specialized POS software compatible than pay for Windows licenses.

> And now it's finally getting close to actual Windows

Security probably hasn't been tested very much but it probably will be as secure as most other OSes? (i.e., not very secure)

What I like about this project, besides the dedication and professionalism of the team behind it, is that it offers PC gamers a possible way to play Windows-era games moving forward. In a decade or two, we likely won't be able to activate a Windows XP machine (and wouldn't want to with all of the security hazards either) so it's great to have a free and open source project to run our Windows stuff.

Recently I had an issue activating XP on Eee PC after a RAM/SSD upgrade (setting it up as a retro-gaming console, better performance than RetroPie). I had to resort to registry hacks and msoobe to suppress activation forever, despite having product key at the back of the netbook - XP CD wasn't accepting it at all. One more situation where legal owners have to waste more time than pirates... If ReactOS makes it far enough to run on Eee PC, I'll get rid of XP.

I use a lot of virtual machines across all versions of Windows. What I've found is that almost every time online activation fails, if I do the telephone activation, it goes though. Just wanted to pass that along in case you or someone else runs into this again.

I tried phone activation as well, it was rejected. I typed all installation numbers correctly.

Dang, I was hoping to help :(

If the automated telephone doesn't work, you used to be able to stay on the line (or just don't enter anything) and talk to a person. They can usually help.

It sounds like you were trying to activate the OEM key on your netbook using retail media.

Yes, that's the case. Unfortunately, ASUS didn't provide installation media, just recovery media, which only works for their 4GB/12GB SSD combo and was failing operation when I tried it on a new SSD. So I tried regular XP Home ISO, but was out of luck.

I'm not a Windows licensing expert, but this is basically how it works:

1. You can't use the OEM license with retail installation media.

2. You don't get OEM installation media because the license is tied to the hardware.

3. The OEM recovery media can't be restored to different hardware.

4. The OEM license key can't be used to install a retail Windows install on a different computer.

It's a strange situation when a Windows licensing expert is actually a thing.

If you think that's bad, you can get Microsoft Certified Professional certified on licensing. On at least 3 separate exams. I am certified on licensing.

Commercial software licensing has always been a PITA, especially at the enterprise level. I just thank sweet baby Jesus every day I don't have to deal with IBM or Oracle.

Simply wonderful! ReactOS is really needed...

I feel you. It really sucks that ASUS's recovery media doesn't let you replace or upgrade your memory or storage. Those are user-serviceable parts, IMO.

It is possible to create an OEM install disc.

I've done it once in the past and it wasn't super complicated. It looks like it might be harder to find the oembios files now, but I bet you can still accomplish this if you're determined.

https://superuser.com/questions/539714/windows-all-oem-activ... has some details that could get your pointed in the right direction

>Eee PC

Now that's a name I've not heard in a long time...

I was using an EEE 1001px as my main computer until a few months ago when I finally got myself a new laptop. I did replace the battery two years ago, but besides that it all parts were the original from 2010. I really like this little computer.

I'm currently using an Asus VivoBook E200H ( https://www.asus.com/uk/Laptops/ASUS-Vivobook-E200HA/ ) as my main computer, which is somewhere between a netbook and a notebook. If it didn't have such limited internal storage (32GB SSD, not upgradable without desoldering) I'd be happy to recommend it. There are plenty of pluses (super portable, long battery life, decent keyboard, etc...).

Microsoft held back the netbook market by setting strict requirements on the tech specs required for Windows licencing, without that restriction I'd suggest we'd still have a healthy market for netbooks today.

https://www.computerworld.com/article/2526983/microsoft-wind...

> ""Our license tells you what a netbook is," said Ballmer at the Microsoft-hosted day with Wall Street analysts. "Our license says it's got to have a super-small screen, which means it probably has a super-small keyboard, and it has to have a certain processor and blah, blah, blah, blah, blah.""

> "Last May, the Malaysian Web site TechARP.com, which regularly leaks information provided to computer makers by Microsoft, reported that the company would restrict Starter to specific netbook configurations. According to TechARP, Microsoft will only sell Starter to OEMs for use on netbooks that have a 10.2-in. or smaller screen, no more than 1GB of memory, a hard disk drive of 250GB or less (or a solid-state drive no larger than 64GB) and a single-core processor no faster than 2GHz."

I was doing "spring cleaning" and found a box with Eee 901, so I was thinking whether to post it on eBay and sell for scraps, or upgrade it and make it a retro-gaming monster ;-)

Funny.... I also just dug an Eee 901 out of a box (it used to run Windows Home Server).

Honestly, just grab a Raspberry Pi for your retro gaming console. It almost certainly does the job better, and the power usage difference will pay for itself soon enough.

RPi is way slower than Eee 901... You'd get many more demanding games working smoothly on Eee than on any RPi. I have RPi as well and was considering making a RetroPie console, but performance was not there. You can even buy faster handheld consoles from China for $25 these days...

I'd suggest a Raspberry Pi 3B+ (can you even buy an original Pi?), with more than 3x the single-threaded performance of an original Pi. I'd wager comparable single-threaded raw performance to the N270, a 10 year old chip at this point.

GPU performance, I'd put my money on the Pi 3B+ over the Eee. Heck, I may even get mine running again to run some benchmarks... or Google for the inevitable raft of folks who already have.

I would need to see data supporting the assertion that it is "way slower than Eee 901". Off the top of my head, I'd expect it to trade blows with the N270. I'd also be surprised if the support for the Eee 901 was anywhere near as strong as the Pi from the retro gaming community, but, honestly, I haven't put together a retro-gaming rig since, maybe, 2001. Maybe it's time to put together two...

I think by the time the OS is usable the available hardware won't have any drivers. I think it will be a VM only OS.

I also have an EeePC, I created an OEM CD looking for OEMBIOS.bin and SLP activation.

I don't remember if it works/worked with XP, but OEM Win 7 can usually be activated via phone, even if the key was refused during installation.

I've been fiddling around with my Eee PC 701 in the past couple of weeks. Is there still a community for this sort of thing? RetroPie supports Eee PC?

Those were fun little machines. I wish mine would still boot... my black 700 has been bricked for years, but it was awesome for college. Especially for those awful, tiny, half-desks that college classrooms seem to always have.

I have the next gen 901; you can use any of the 32-bit emulation platforms on Windows and the 1.6GHz Atom (that could be overclocked to 1.8) is significantly faster than any Raspberry Pi, so you have a larger pool of games you can play at a proper speed. I upgraded it to 2GB RAM/60GB SSD, and it really flies. It's only slightly larger than a handheld console, so I thought converting it to a retro machine would be perfect use for it ;-) I tried Linux Mint on it initially but it was much slower than XP, so I am sticking with XP.

The last time I was using my eeePC 901, I managed to put Windows 8 on it, and I remember it being faster than either the XP or Ubuntu/Mint varieties I tried on it.

Slitaz would blew both 8 and Mint. Try it.

I bricked mine ages ago, unfortunately - I really liked it, despite the keyboard being just a bit too small.

Yeah, DOSBox is nice for retro gaming or even the oddball application, but it pales in comparison to the potential of this project.

The place where DOSBox isn't sufficient is for games in that Win 95/98/2000 era. Most of those don't run very well, if at all on modern Windows, unless somebody still has the source and bothers to rerelease a modern "HD edition" or somesuch. I still have to keep a Win 98 VM so I can play Sierra's Civil War Generals 2 from time to time.

I heard that DOSBox can run Windows 3.1 to run 16 but Windows games. I also heard someone hacked Win95 and Win98 to run under DOSbox. You could just use QEMU emulate a Pentium box and install 95 or 98 under that as an alternative.

Dosbox runs windows 3.1 perfectly, with resolutions up to at least 1280x1024 (what I use for tetris) _or_ true color mode (unfortunately not both). You can supposedly install 95/98 in it, but I've had trouble actually using them in it.

Cracks will always be around... and XP is one of the "simpler" ones compared to the newer Windows.

"security" isn't a big deal either, just keep all your ports closed and don't run untrusted code or unnecessary services. AFAIK all the exploits have been on services and such, which shouldn't be exposed to the Internet anyway.

Life gets a lot better when you start running Win10 security.

IE: Win10 can run the password manager service on a separate VM now automatically. Hell, you can start new version of Microsoft Edge in a clean and isolated VM, requiring the attacker to use a hypervisor zero-day to pwn your box. And since the Microsoft Hyper-V hypervisor is often running through UEFI secure boot with assistance of the TPM modules on the motherboard, and because the Hypervisor has very few services running (really: any service is in its own VM), its a much, much harder attack surface to go through.

https://blogs.windows.com/msedgedev/2016/09/27/application-g...

Security features that add convenience (ie: group policy allows admins to automatically open websites in "Application Guard" mode) helps a lot. When the user literally only has to wait ~5 to ~10 seconds to boot up a clean version of IE isolated inside of a VM (separated in a RDP session), its much easier to have widespread security throughout a network.

I mean really: think about the security model needed to pwn AppGuard + IE. You either need to pwn the RDP session (unlikely, but hey its possible). Or, you figure out how to escape an isolated VM, when said VM has virtually all applications locked out. I saw a security demo of AppGuard recently. You can't run cmd.exe, you can't run powershell, you have zero permissions inside of your VM. If you somehow escape the sandbox, you STILL have to break through a hypervisor to get to the user.

Its way, way WAY easier to do "proper security" with Win10 + all of the virtualization tricks (AppGuard, Credential Guard, Device Guard). The modern Win10 security model is beginning to be immune to even kernel-mode exploits.

-----------------------

With that being said: WinXP has security problems with modern Win Vista+ because that's the edition where Microsoft decided that direct-hardware access is a BAD IDEA for standard usermode applications.

WinXP allows any user-mode application to directly talk with the hardware. Win Vista+ does NOT allow it. And that broke a ton of programs (old controller hardware, printer drivers, etc. etc.)

Fixing security issues causes compatibility problems.

Newer Windows is still only two shell commands away from activation.

Even better buy a $10 key on ebay.

Do tell

KMS servers. there are a bunch on the internet you can use, and you can even run one locally on your LAN. only catch is you can't run the server locally (although you can patch out system files to bypass this), and you need to connect to the server every 180 days to refresh your ticket.

Ive considered this in a few things.

What if we could go back and access that infrastructure, what lessons could we learn?

I really dont know, but I personally see this as useful from an entertainment POV, let alone function.

Good thought, never saw it from an archiving point of view!

Don't all new Windows releases ship with backwards compatibility modes?

It's already quite well known that some Windows XP era games cannot run properly on Windows10. It does not affect every game out there, but it's easy to find enough to say this is a problem.

I wouldn’t count on Microsoft supporting XP era software 10 or 20 years from now, and I wouldn’t ask them to.

I recall seeing videos on YouTube of Windows 1.0 applications running on Windows 10, so I suppose it really depends on what type of software exactly; more system-level things like disk defragmenters are unlikely to work, but basic productivity software likely will continue working. I personally have a few from Win95 era which I still use regularly, and of course I'd expect the basic Hello World messagebox binary to remain usable 10 years from now.

I doubt it will be too long before 32-bit versions of Windows are no longer produced or supported. The 64-bit versions do not run 16-bit software. When that time inevitably comes, the Windows 1.0 software will also stop working in native Windows (rather than a 32-bit or even 16-bit Windows version running through emulation).

Basic applications such as that, sure. But the top parent comment referred to games, which are the opposite in terms of complexity and operating system features, and generally harder to support adequately in emulation.

They do until Microsoft decides they don't. Unfortunately I don't know of any guarantees that MS has made about forever supporting XP comparability.

I feel like ReactOS has gained more attention lately and hope some of it will spill over to support of this project. It’s a bit like you need to reach a critical mass in features and usability, and until then it’s an uphill battle to maintain enough traction. Haiku is another cool project, especially as someone who will always have a warm spot left for the Amiga. But until the long waited for beta is out, I think they have an ever more dangerously steep uphill path to climb.

Haiku OS[0] is a BeOS inspired OS, though I do see some articles about a port of it to Amiga hardware.

[0]https://www.haiku-os.org/ [1]http://www.generationamiga.com/2017/09/03/haiku-os-for-amiga...

They must've meant AROS http://www.aros.org/

Within the Amiga community there's a certain level of admiration for HaikuOS, as BeOS/HaikuOS is seen by these individuals as a spiritual successor to AmigaOS. As you point out, AROS is the open-source reimplementation of AmigaOS, so it's worth mentioning it, I just wanted to clarify why the Amiga+Haiku link might have been made.

Much appreciated! I wasn’t lucky enough to have an Amiga, I only had a PCjr. I did see them in the store and listed after them.

Those of you discussing Amiga might find this comment from skrzyp interesting:

https://lobste.rs/s/vkpekg/morphos_3_10_released#c_ktgnn3

It gives a huge, deep history lesson from author's perspective covering important events, hardware/software, and especially the culture. It was fascinating. Also made me hold off on MorphOS for one of my projects.

Thanks for sharing the comment, but it's got a number of inaccuracies. For example...

> "The problem was that OS wasn’t completely prepared for that, but clever people from Phase5 found a way - they made their own microkernel (called PowerUP, later WarpUP)"

PowerUP and WarpUP were competitors, and were the seeds for the later Red vs. Blue civil war.

If you'd like a clearer history of the Amiga, I'd recommend this series of articles on Ars Technica (it doesn't mention AROS, but is otherwise a good introduction):

https://arstechnica.com/series/history-of-the-amiga/

One day when Windows drifts too far away from its golden age functionality, it will be great if ReactOS can fill the gap. However, I feel like MS could one-up them overnight by providing a sandboxed 32-bit XP VM as part of current windows. We’d get back DOS support and compatibility with old programs would be better. Call it Windows Subsystem for Windows or whatever. That would also pave the way to removing a lot of Win32 cruft from the main codebase, which MS seems very itchy to do.

It was already a thing in Windows Vista - at least in Enterprise, you could download a "windows xp " compatibility package, right click on any application and start it in windows XP - it would silently start XP in the background and render the app the same way as a native Vista app, while running on the XP kernel in the background. I have no idea if it exists in newer versions of Windows however.

No it doesn't, unfortunately.

The best option they suggest is moving it to a VM under Windows 10's Hyper-V:

https://social.technet.microsoft.com/Forums/windows/en-US/1f...

What's the point in doing this for MS? There's no money in it and they'd need a serious team backing it. And by no money, I mean at scale - there are going to be small companies needing something like that and there are going to be huge companies. Huge ones already got a public offer of "pay an increasing tax for XP support" and likely a private "this corporate deal is possible" if they're important enough. The small ones put together, I don't think could collectively fund ongoing xp support.

If anything ReactOS picking up those small companies is beneficial to MS by keeping them in the MS ecosystem.

Like a sibling comment mentioned, it wouldn’t be the first time MS has done this sort of thing (XP on certain SKUs of Vista or maybe it was 7). I assume they already have a serious team backing tons of legacy Win32 cruft. So, instead of maintaining that, they freeze it and shift it into a VM. MS gets to cut down modern Win32 to something closer to the UWP subset, and users that want it—-make it an optional subsystem—-get better legacy compatibility than they have had in years.

Maybe the 2020 win7 end of life would be a good time to do something like this. I don’t know if there is a good enough financial reason for MS, but from a user’s perspective I think it would be better than the situation today.

> Several bugs have been squashed in the ReactOS CC and Freeloader, allowing it to boot in 96MB hardware (smaller RAM need than our 0.4.7 release).

This is an impressive update when most projects, and not necessarily wrongly, always move in the opposite direction. Good to see.

As a free Unix user I'm primarily interested in ReactOS in order to test certain Windows things such as cwrsync over ssh. I tried installing ReactOS about a month ago but didn't have the right hardware to make it through the installer. I will definitely try again.

Windows 10 is still not a good option for embedded x86 systems even in its proper supported embedded mode. Once I've replaced the shell, turned off all the services and junk programs and opened the one port on the firewall my industrial touchscreen application requires I would really rather windows take the hint not to change anything and leave things be.

I have been keeping an eye on Reactos for over 10 years, it is still a promising candidate for my legacy systems, USB Install would still be good to have as it is painfully hard to get it on to newer systems which don't have Optical drives.

It took me a while but I can now compile my gui applications for Linux, so going forward that is my strategy, but lots of factories out there still scared to move off windows XP and Reactos would be ideal.

Seen ReactOS pop up a few times and always thought it was something to do with the Facebook Javascript library. This is really amazing, I'm looking forward to seeing where it goes.

Yeah, I've been confused too. The website itself doesn't help: no "About" or "What is ReactOS?" in the main nav. I ended up looking it up on Wikipedia to find out what it is: an open source OS that is intended to be binary compatible with Windows.

This does actually ring a bell for me - I seem to remember mention of it very early in the century. That perhaps cuts to the heart of the issue: it's been going as a project since 1996, yet it's still only considered alpha. In other words, I'm not so sure it's going anywhere fast, and I'm not sure what you'd really use it for.

I bet in 20 years no one will know what ReactJS even is, but ReactOS very much will preserve NT regardless of Microsoft’s successes.

ReactOS is a usable Windows NT 6+ clone, with kernel ABI ( driver ) compat, that runs real software. What are you comparing it to, to be able to qualify the speed of its development?

> I bet in 20 years no one will know what ReactJS even is, but ReactOS very much will preserve NT regardless of Microsoft’s successes.

Don't be sure about something in the tech on such a long time frame. Just to put things into context. Andoird is less than 10 years old and already close to windows market share territory. Facebook is only 14 years old.

I can't even imagine to comprehend what will happen 20 years from now.

The JS ecosystem is as of now still a rather ... fast paced environment. There is a lot of churn in there. I don't think ReactJS will survive ReactOS.

There are hundreds of millions of lines of ReactJS in production today and hundreds of millions more coming. If you think 20 years from now all that is going to be re-written I've got a bridge to sell you.

JQuery is 11 years old now. ReactOS is a goal and nothing more. Nobody is using it right now because it's barely usable outside of a VM, and even then it's software compatibility is smaller than WINE, let alone Windows.

ReactOS is a cool project, but your comments are bordering on delusional.

> There are hundreds of millions of lines of ReactJS in production today and hundreds of millions more coming.

10 years ago there were 0. The churn in JS code is just too high.

10 years ago was JQuery, Backbone and KnockoutJS. There are still millions of lines of all three of those in production today. The churn in popularity is high, but there are MOUNTAINS of legacy code out there.

I was surprised to find it has been going since 96. There is some good info on this site: https://www.linuxinsider.com/story/83578.html

Blame Facebook for poor name choice. ReactOS is much older than React, the JS library.

Lots of products have similar names, that doesn't make ReactJS a poor choice. One is a frontend framework and one is an operating system. The naming isn't the reason he wasn't aware of ReactOS... the fact that it's a very small community is why he wasn't aware of it.

I realise that now but did have to do some googling to find out about the history of the project.

I didn't downvote you (seems a bit harsh), but just so you know ReactOS development has been going for over 20 years, which is a lot longer than ReactJS (which is approximately 7 years old).

I expected to be downvoted for not knowing this.

I think that this comment has been downvoted is highly unfair.

With the US Court of Appeals for the Federal Circuit ruling that API's are copyrightable I wonder what type of impact it could have on open source projects, such as ReactOS, if the copyright holder ever decided to take action.

1. The ReactOS project's implementation of Windows APIs is purely for compatibility reasons; on the other hand, Android's implementation was not meant as a compatibility layer for existing apps. (In fact, that Android didn't implement a compatibility layer was a huge sticking point for the Sun contingent within Oracle—although if we're honest, we know the motivations for Oracle proper are the dollar signs.) It wouldn't be a good idea to try and draw direct parallels between the two.

2. Oracle v Google still isn't over, so even drawing any inferences about Oracle v Google itself would be premature.

3. The copyright holder here is MS. Any action on their part to go after ReactOS now be the undoing of all the goodwill they've built in the Nadella/MS-on-GitHub era. The result would be (a) a bunch of I-told-you-sos from the leery folks still holding on to their grudges today, and (b) a massive, massive, innoculating "fool me once…" reaction among the folks who'd actually been made suckers for not heeding the grudgeholders' warnings. All in all, it would be a spectacularly bad idea for MS at this point, especially given how much weaker their empire has already gotten in the last decade.

One of the most interesting changelog that I have read in a While. Well written, thoughtful and gave me good insight into the project.

Indeed a great read to catch up with the project.

> Talking about the notification tray, due to Ged’s work, icons of killed and finished process are now automatically removed, even when apps crash. This is something that Windows doesn't even provide with Win10, and many Windows users may have noticed.

Made me chuckle. This has bothered me since forever. It seemed like such a low hanging fruit to fix, I wonder why Microsoft never did.

> I wonder why Microsoft never did.

I'm assuming it's because of the size of the codebase, available engineering resources, return-on-investment calculation, and a business prioritization of new features over refinement.

Most engineers can look at their own work or codebase and think, "well, that's a bit shit but it works well enough." I can't imagine their backlog. I'm sure it's lurking in an issue queue somewhere.

One notable thing is that in Windows the behavior of taskbar changed around the time of Windows 98SE/2000 (I'm not sure in which version exactly). Originally the "orphaned" tray icons just were there and you could not cause them to go away, "new" behavior is that first event (ie. hover) causes them to be destroyed.

Guesswork: once you hover over an icon in the taskbar, later Windows checks if the process (or, rather, window, as the icon will be tied to the window handle you get from CreateWindowEx) handle belonging to it is still alive. If not (aka it cannot find a target window to send the message to) the icon gets destroyed.

ReactOS, another guess here, does either a polling check if all taskbar icons have valid process handles or a check whenever a process exits.

AFAIK the taskbar forwards the window messages destined for the icon to window that owns the icon and in the "new" Windows' behavior simply notices that such window does not exist. ReactOS behavior probably works on the basis of hooking destruction of the owner window (although I suspect that on Windows windows owned by killed process will not get WM_DESTROY message, but it's wild guess as I'm completely out of my rudimentary and rusty knowledge of WinAPI there).

The notification area API sucks because of considerations made 20 years ago. The icons aren’t windows, the application just sends bitmaps to show and is manually notified if you hover over or click on an icon.

Explorer can’t tell if the app connected to an icon crashed without polling or other tricks which Microsoft probably wants to avoid. But it can tell if you move the mouse over them and it can’t send the messages to the application.

Also the notification area with a million icons is bad UI that I can imagine Microsoft not wanting to invest in. They tried to discourage its use by hiding most of the icons all the time.

The reason could be stated in one word: "architecture", or perhaps less positively, "bureaucracy". In other words, either they could've designed it originally in such a way as to make fixing such a bug very difficult to do without rewriting massive amounts of code, or the fix could be simple yet the politics around actually getting that fix into code are not.

Having worked at large corporations (not Microsoft) before, I can definitely see how things like this happen --- and the reason why "enterprise" software tends to have a lot of these superficially simple and annoying defects. To add insult to injury, the codebase is often offensively overengineered and in precisely the wrong direction to facilitate the change required to fix such bugs, and even the tiniest of changes requires a ton of extra paperwork, approvals, and reviews.

Vista and 7 had a well known bug where a contextual window would stay on screen forever, until you changed the resolution or logged out - I think Microsoft commented on this officially, saying that fixing it would need to be done somewhere low at the kernel level and it's simply too much work. I think it was resolved fully in 10.

I wish the Gnome release notes were more detailed, like this one.

I wish all release notes were as detailed as these... Now a days you're lucky if you get anything aside from "Bug fixes and minor enhancements" with some releases... What changed??? It's useless to have a changelog if it doesn't document anything substantial.

The What's new blurb in the Play store is regularly panned for this. Every big name app, even Google's own, constantly read "bug fixes and performance improvmements"...

I've submitted non-clean-room reverse engineered code to ReactOS years ago when I was younger and it was accepted with no questions asked. But since that day, I can't bring myself to believe all of this is clean-room reverse engineered.

I hope it was an outlier event and that external contributions are better audited today.

As long as nothing can't be proved, why bother? I don't think one could raise a good point about ethics in favor of Microsoft. DMCA is shit.

Superman flies. Spiderman spiders. Trolls exist.

Genuine question: What advantage does ReactOS provide over Wine?

https://www.reactos.org/wiki/WINE explains a bit but still doesn't answer the question.

P.S. oh by the way, the developers of both projects are apparently collaborating and ReactOS uses Wine's DLLs for their userspace, so my question is not meant to criticise the "double-effort" or anything like that.

The big-picture advantage is that Windows users can switch to it without becoming Linux users first. Many people actually like (some version of) Windows and would prefer to keep using it, but have issues with licensing or Microsoft's stewardship of the platform.

There are also some technical advantages for applications that require driver code. Examples include specialized software with intrusive DRM (where either the DRM is so beastly or interest in the software is so narrow that there's no crack; ISTR that Cubase is sort of legendary for this) or software that drives industrial equipment. Wine will probably never run this sort of thing.

Well, for one, you don't need to deal with setting up a separate Linux OS.

Primarily support for drivers, I think.

https://wiki.winehq.org/Wine_History and https://www.reactos.org/wiki/ReactOS/History were interesting to look through.

I've wondered this myself, anyone know if ReactOS can run anything wine can't?

I tried wine-develolment a few days ago and 3D performance is still rather spotty (amongst all the other usuall hiccups). That being said is has helped me play a few games with a decent experience.

Drivers for one - reactos can be used for Windows driver testing, wine cannot.

I don't know for you but wine never really worked for me. For any app that I would like to use, there is some quirk, after a week of trying I gave up on skype for bussiness and rather run a virtualbox just to be able to use it.

The USB support isn't quite there yet (major rework didn't make it this release) but otherwise this is a pretty good release with massive progress in actual usefulness.

For why USB is hard, see http://www.usbmadesimple.co.uk/.

Ugh. This website saved my ass trying to make HID devices work on top of the completely non-conforming Linaro DwUsbHostDxe driver (look at https://github.com/andreiw/RaspberryPiPkg/commits/master/Dri... for a laugh). I did end up having to hit the USB spec in the end, but it's a legit resource to get a birds-eye view of the whole USB1 and USB2 thing. It doesn't make USB simpler, but sure makes it more approachable.

Offtopic: What's the end goal of all this UEFI work you're doing for the Pi?

No concrete goal beyond having a reasonably useful UEFI development environment (more so than the 96boards one).

It’s been useful to be able to run the same Ubuntu and SUSE images as VMs and servers do, but that’s only really possible because the built kernels have RPi3 support (but at least you now don’t need a custom image with U-Boot).

Actually, good question to flip back to you.

As a random person stumbling upon this, what value would you assign to it? What could you imagine the end goal to be?

Arg! This is so awesome, but is as though USB 3 never existed.

Tried the livecd in a VM and a couple of niggles sprang up.

First of all, no driver for the network card in Virtualbox, though Vmware worked fine.

Second, no web browser...

Third, the mouse sometimes freaked out and got stuck in a corner.

It should be obvious enough, but perhaps you missed: you can install a browser, such as an old version of opera, by fetching its installer over the internet via ReactOS builtin "program downloader" (I do not remeber its name).

If you meant lack of default browser preinstalled though, I agree with you its lacking.

It's necessary to change the network card in VirtualBox for internet connection to work IIRC.

Ah, i missed the program downloader while poking around.

I'll take another look.

Firefox 48 worked, but the autoupdate crashed.

If you install 7Zip, you should know that 7Zip does not associate itself with file extensions by default. This is 7zip behavior. You need to run 7zip, go to options, and set associations by clicking on the '+' button.

AMD PCnet-PCI II as the NIC works fine for me (VBox 5.2.4 r119785 (Qt5.6.3) on the Mac, if that helps)

I’m curious - so now that the ReactOS guys are picking up bugs in Windows, does this mean they have a comprehensive test suite of Microsoft Windows proper?

They've had for a decade at least.

I'm willing to bet, this will be very useful in a few years time, and the primary use will be 'fixed function' devices (in industries that are very slow to change) that currently require Windows. (ATMs, POS, Announcement Screens, Equipment controllers, SCADA systems, etc, etc)

I think their site is getting the HN hug of death, loads extremely slow/doesn't load at all...

yep it's currently fully smothered for me

I think Windows will just keep breaking compatibility faster then they themselves and others are able to keep up. Forcing users to keep upgrading and buying the same software over and over again.

I wonder if the planned Vista/7/10 support will extend to kernel mode drivers, especially for WDDM2 and up graphics drivers.

Currently, the supported graphics drivers is pretty sad for modern-ish hardware on ReactOS, because they are limited to drivers for Windows Server 2003 / Windows XP, which aren't available for a lot of modern cards (or integrated chips, for that matter). Additionally, a lot of the older XP drivers don't work well on ReactOS today.

I love ReactOS, especially in VM. Windows requires separate license for each VM. If I maintains 10 VM for testing, it's a problem.

In case you are dismissing this as a fringe project, remember, what is ridiculous today is commonplace tomorrow.

I'd be curious if ReactOS runs the Oculus Rift drivers, runs docker and is more performant then Windows 10.

> more performant

Do you mean faster?

More resource fficient, which ususally means faster in execution or runs better on resource constrained machines. For example how well does it run in 2G compared to Windows 10?

Can anyone comment if I can install old Windows 2000 games on this, like C&C Red Alert? Would it work?

Hit and miss as there are missing win32 pieces, but Red Alert for example ran nearly as well as it used to, for me.

PSA:

The best way it seems to check the AppDB is to do something like

  site:https://appdb.winehq.org "simtower"

and

  site:https://appdb.winehq.org "c&c" "red alert" "2"

(The search box on the website is just google custom search)

C&C Red Alert 2, both 1.0 and Yuri's Revenge: rating "everything works": https://appdb.winehq.org/objectManager.php?sClass=version&iI...

And simtower?

- Version 1.0: Game saving is broken, everything else works: https://appdb.winehq.org/objectManager.php?sClass=version&iI...

- Version 1.1b (retail): Copy a DLL over for working sound, then everything works: https://appdb.winehq.org/objectManager.php?sClass=version&iI...

What Kernel does this OS use? What's its history? Is it a unix fork such as Linux?

No, it's a reverse engineered Windows. I believe that due to copyright issues it even is a fully clean-room developed implementation, which means that it's not based on MS documentation or on decompiled source code.

> which means that it's not based on MS documentation

I think looking at MS documentation is okay. You're allowed to look at MSDN for Wine development[1]. Since the projects share code, you're probably allowed to do so for ReactOS as well.

[1]: https://wiki.winehq.org/Clean_Room_Guidelines

The ReactOS kernel is their own, clean room reverse engineered to be compatible with the NT kernel.

Also just a side note, Linux isn't exactly a UNIX fork either. Linux was developed from scratch on a MINIX host, which in turn was another OS designed to be UNIX-compatible but did not actually use any UNIX code. To put it another way, it's Unix-like two generations removed, but definitely not something you could call a fork. That might seem like a small detail to you but SCO vs IBM says otherwise.

It uses it's own opensource implementation of Windows NT kernel.

And a lot of code from Wine too

See eg: https://github.com/reactos/reactos/blob/master/ntoskrnl/incl...

appears to work well, including old photoshop 5.5 from a windows cd https://youtu.be/Ne88Is2cymQ

Does Starcraft run?

So I'm a ReactOS fanboi, but... but...

- You need to fix your flaccid dev community. Your forums have a terrible S/N ratio. IRC may be great and all, but I don't see any hard documentation behind any design discussions or development. I do see crazy shite like "can I run ReactOS on Playstation 2". The forums make your project seem dead, which it clearly isn't.

- You need to fix your flaccid dev community. It needs a larger presence. It needs to be tangible (who is doing what? who are the people? do you have conferences?)

- You need to fix your flaccid dev community. There is a bunch of really useless, out-of-date, confusing and out-right self-inconsistent information on building ROS, and the Git move didn't help (but kudos on the move, it is long overdue). I've tried 4-5 times to do something, but the directions I followed on Win, Ros or Linux all ended up breaking somewhere.

- You need to fix your attitude problems (or your forums). There are people asking reasonable questions, who get actively or passively aggressive answers. Not good.

- You need to fix your toolchain problems. This whole SEH thing is getting out of hand. Does Clang support SEH yet? It might - https://clang.llvm.org/docs/MSVCCompatibility.html. So maybe you need to bail on GCC, or maintain your own fork of it (yes, it's ridiculous that GCC still doesn't have SEH support, and the patents have long expired).

- You need to get a grip on x64 support. Your toolchain problems cannot be a gate (and you can clearly use MSVC for building... so...). IA32 support is great and all, but needs a back seat to x64.

- You need to get a grip on UEFI booting (note UEFI booting does not /necessarily/ mean you need improved ACPI support, in practice)

- SMP?

- Forget Xbox, NewWorld PowerMacs (those were BE anyway) and 32-bit Arm chips (which outside of the Pi are too varied and too few). Targeting 64-bit Arm chips does make sense, though, now that Microsoft has released a client build of Win 10 on laptops and an Arm server vendor demoed Windows Server at OCP'18 (http://www.opencompute.org/assets/Uploads/18150J-Ampere-PPT-...). You will need UEFI and reasonable ACPI support to support ARM64. You can target SBSA and SBBR compliant systems and this will give support for the entire server ecosystem (and qemu VMs, har har).

- You might wish to minkernel-ize ROS as well, so you can boot without a GUI.

- You might wish to update Wiki (and build a list of links to relevant pages) for active subprojects.

Have you logged bugs on all of those instruction issues you had?

I suspect the core devs for ReactOS are very few in number and this probably gives rise to the issues you see. But a rant here won't fix much.

Wow, aggressive much?

I dunno dude. I just pointed out the obvious, and all I get is a passive-aggressive minusing. Most people won’t care enough to even give you “a rant”, they’ll just move on.

If you don’t want feedback, don’t ask for feedback by posting publicly. My “rant” is a collection of issues I’ve seen for at least the last 7 years of actively paying attention to the project.

Every time I see this I briefly think it's a Linux distro with a React-based window manager

Reply on news.ycombinator.com