Submitted by Oh ReactOS. This project has been with us for a very long time now, and since day one, I never really knew what to think of it. They always seem to be running at least 300 miles behind the Win32 bandwagon, but what they’ve accomplished so far is insanely impressive nonetheless. This new release comes with quite a lot of new stuff.
First and foremost, there’s a new networking stack – or, more accurately, lwIP has been ported to ReactOS. Both stability as well as performance has been improved this way, so that even things like torrents can download just fine now. Compatibility with NT 5.1 drivers has been improved, and it also supports wireless networks, both open and WEP.
Users can now also install and use XP-based themes to prettify their ReactOS experience – or simply revert to the classic theme, of course. In addition, thanks to the addition of a kernel mode testing framework, several issues regarding kernel mode components have been identified and fixed, improving core stability.
Another major change is this: “Support for building ReactOS using the Microsoft toolset has also played a large part in the release. This is now at a stage where ReactOS can be entirely built using an MSVC based environment and can produce a working boot and livecd. Along with GCC, ReactOS can now be built using two different toolsets and on a variety of platforms including Windows, Linux and Mac OS X.”
…although to be fair, I don’t expect much. I’ll play around with it a bit, but I’m sure the fun will be short. It always is–because I always very quickly run into crashes. Last time I used it, it felt like it was approaching the stability of Win9x (which itself was not that good). I’m hoping these major bugs get squashed sooner or later… better sooner.
Around the 2005 or 2006 I used to have a fairly decent P3 laptop it worked great on. Well great as long as you didn’t expect more than a Windows 95 level of experience.
Why is it that when you say even the slightest criticism of an open-source project, you always seem destined to be modded down?
It’s not what you said. It’s how you said it. Saying that the developers need to address stability is one thing. Comparing its stability to Windows 95 is easily (mis)interpreted as trolling.
Bleh. I guess comparing to another OS that it is is supposed to be mimicking as close to 100% as possible is somehow “wrong” then? Where the f*** is the logic in that? Maybe I should be comparing ReactOS to Mac OS 9 then? Or Linux? Or maybe even one of the BSDs? Then maybe I’d be modded down for trying to compare it to unlike operating systems. WTF? I mean… come on, seriously.
Yeah, yeah–you can say that “these days ReactOS is targeting Win2000”. But I can say right back, sure, but it was originally targeting Win9x… and it has yet to get caught up fully to *either* target. No matter what Windows version it is targeting, it is Win32, which means both Win9x and WinNT (32-bit).
I say that it has has yet to surpass its predecessor from 1995 in terms of stability, what’s so unfair about that? Seriously… is ReactOS trying to compete against something besides Win32 these days or something?
Windows 9x has a totally different architecture then the NT family. So just saying it’s Win32 is not really working. It would even be more adequate to compare it with NT 4.0 then Windows 95.
Well maybe I should’ve clarified myself by saying “Win32 API” which runs on both of the OSes that I mentioned…
At least we know where the Start Button will live on, right? (vs. http://www.osnews.com/story/25581/Microsoft_Removes_Start_Menu_Butt… ) And no Metro.
Sometimes I wonder what could happen if ReactOS were to be ~ready when corporate XP machines start to be decommissioned on a really massive scale*…
…nah, such places would probably at most run XP under virtualisation, I guess.
(* or maybe it is already happening, already a bit too late)
Yep. Several of our customers have been investigating that approach – Windows 7 and IE8/9 desktops for those apps that will certify on a modern desktop, and VMs running XP and IE6 for the legacy stuff.
Those people are never going to look at ReactOS, because even if it were perfectly compatible, they’d have no support from vendors.
Nice seeing a new release, some really good stuff in there too. For such a small amount of developers it’s a really impressive achievement that they’ve gotten the project this far.
Well it could be a business opportunity for the ReactOS devs, however if ReactOS would reach a point where it could be used commercially as an alternative to say XP then Microsoft would hit it with the patent hammer faster than you can say -‘oh shi..’ .
The good thing is, my understanding is that ReactOS is based in Russia.
So, it’ll be significantly harder for Microsoft to sue, unless Microsoft sues ReactOS users (which, well… they’ve shown a willingness to do that before – see the whole SCO fiasco – as the old Slashdot troll went, “don’t forget to pay your $699 licensing fees, you cock-smoking teabaggers!”)
I remember a story a while back where ReactOS lobbied the Russian government for funding. I wonder how this panned out.
http://www.reactos.org/ca/news_page_67.html
Always glad to see some news from the alternative operating systems category. It feels like the trickle of news there is slowly drying out.
Congratulations to the ReactOS team on their release!
I played around with the live image in QEMU, and as always, their work is impressive.
Edited 2012-02-08 00:35 UTC
I always felt that if they could get the LiveCD portion working well enough, and had a way to build custom LiveCD, it would make for an excellent and legitimate alternative to WinPE, BartPE.
Don’t think I’ll ever need it personally but the progress is great.
This could be used by bank atm’s and heaps of other devices which still run on windows xp.
I am impressed. Since the last time I saw it (probably years ago) the progress has been remarkable.
I wish I could really use it instead of Windows, but I use Windows only because of complex apps not available on a Mac. Thus I am afraid it is an impossible dream.
Edited 2012-02-08 05:33 UTC
Haven’t tried ReactOS in a VM or Wine http://winebottler.kronenberg.org
http://wiki.winehq.org/MacOSX
or Play On Mac http://playonmac.com/en/ to get your Windows apps to run without using Windows directly?
They are very complex applications, as I said. Some of them will run in Parallels.
WINE and all its by-products are doomed as long as they depend on X11. X11 on Mac OS X never showed much potential and hasn’t changed in ten years, so to speak.
There isn’t a strong WINE community on Mac OS X — X11 was a temporary evil for running it, an implementation using Quartz (Mac OS X native display system) was coming. It was 2006 or 2007. I’m still waiting.
Well, it has changed. It have been replaced by XQuartz and it is based on recent Xorg code. It is also installed by default.
WINE and all its by-products are doomed as long as they depend on X11.
Good that Wayland is well underway then
Seriously, the ReactOS Win32 subsystem is largely Wine and it isn’t running on X11. Wine itself will probably switch to a more modern display system when it becomes sensible to do so.
Spelling doom for a project that is nearing 2 decades of existence is funny though.
Who needs ReactOs? I DO see the need for WINE, and I understand that WINE and ReactOs developers cooperate, which is a good thing.
But I can’t see the need for ReactOs. If you need an OS to run Windows programs, why not use.. eh.. Windows?
Edit: in the longer run, the lack of safety of XP might be a problem. My solution would be to run it in a virtual box and kill that box (including all possible malware) when you’re done.
So: favourite OS > Virtual Machine > XP > programs
Edited 2012-02-08 13:27 UTC
Licensing costs, licensing terms, or disgreements with recent design choices by Microsoft spring to mind immediately. There’s also a significant case to be made for having the ability to modify the OS as needed and having multiple potential venues of support on a timetable that isn’t controlled by a single corporation.
Anyone for whom the licensing costs would be prohibitive will be a large enough organization to afford the licensing costs, especially when that cost is basically built into the price of modern PCs.
ReactOS is to Windows as Linux is to UNIX. It’s that simple.
Not as accurate; maybe fifteen years ago, but not now. Nowadays the leading Unix-like technology is being written in Linux before than being ported to the real UNIXes and BSDs. In several cases, the ports are second-class citizens because not all UNIXes implemented several things built in Linux (udev comes to my mind).
Surely that even more of a reason for the existence on ReactOS? A free software implementation of the win32 and NT API is able able to be tweaked and changed in ways Microsoft wouldn’t even dream.
Plus, a complete cleanroom rewrite of Windows Server 2003 (as ReactOS is “cloning”) is likely to produce a cleaner, more efficient and lightweight system than Microsoft’s own Windows Server 2003, as the codebase that Microsoft has dates back to the days of Microsoft’s OS/2 v3, which NT evolved from.
Maybe, perhaps, but Linux certainly started out as a UNIX clone.
I’m not so sure that’s true. Solaris had dtrace and ZFS before Linux developers responded with SystemTap and Btrfs, neither of which have achieved feature parity yet.
Yes, you are right on this and other technologies being developed in the BSD/Unix fields.
Solaris would like a word with you.
They do a bit more than cooperate. Without ReactOS, you wouldn’t have Wine/WineHq.
The only need I have for Windows is to run proprietary VPN clients required when connecting to our customers’ networks. This will never be possible with WINE. I currently run XP virtualized, but as someone else mentioned, this requires a license (just to connect to a VPN…)
The only need I have for Windows is to run current FOSS products. Addditionally some of them are windows only (miranda/infrarecorder/notepad++/cdrtfe/peazip/7zip) I wish they were cross-platform.
There are linux equivalents, but on windows some of them are really nice.In any case I wouldn’t lose my sleep if windows went down tomorrow. I am usually on my linux desktop but ReactOS is a good FOSS home for this FOSS work.
If Firefox/Libreoffice/Erlang/Sharpdevelop and some other MSVC only FOSS were buidable with mingw on ReactOS, the world would be a better place.
I would have no problem to run ReactOS on a VM (but I need a new HW/ currently on a D510 atom with opensuse 12.1_x86-64)
To answer the question of why:
1) Why not
2) Just because
3) No licensing costs
4) Open source implementation
4a) Not just of another *nix, but a completely different kernel
5) Binary compatibility meaning any existing drivers should ‘just work’
6) Educational value
7) Do I really need to keep going?
Yes, you really should keep going. Unless you were making a joke, that list was not at all convincing.
Why would you come to a site that’s main purpose is to talk about operating systems (alternatives in particular) and then ask why we are talking about an operating system?
I don’t think he really asked that – more like wondered what’s the point of that particular project (which falls under “talking about an OS”), a perfectly valid question and to which answers seem to be… vague.
Anyway, in regards to “a site that’s main purpose is to talk about operating systems (alternatives in particular)” …I don’t know, have you seen the main page lately, how many news are about alternative OS in particular? (note I don’t count Android, or Linux in general, as such)
Indy OS scene seems mostly dead …seems it was a relatively short burst before people gave up / fully accepted the field is largely delineated already / it got boring.
Edited 2012-02-15 00:05 UTC
I’ve been following ReactOS since 0.3 and had high hopes that it could reach a usable state soon. 6 years later and still we’re not there. Thankfully during the same time Wine improved considerably and I do use CrossOver to run some Windows apps now and then.
With the way things are going ReactOS will never replace Windows in any sense, which is a shame because I wanted it to reach that critical point where it can be a drop-in replacement for a segment of the population, not necessarily everyone. ReactOS is an amazing project but I guess somethings are harder than imagined.
I browsed a lot of niche OS sites lately scouring for news to submit.
Some of the reasoning below might be false. Reason some sites mainly update news through mailing lists and not on the site itself.
But here are my general impression.
Seems like the last of the golden OS hobby years ended in 2009.
Jnode
Last news post 01/29/2009
http://www.jnode.org/node/2879
Coyotos
Last news post 16 May 2008
http://www.coyotos.org/
DexOS
No front end news items.
Forum not much help
http://dex.7.forumer.com/
HelenOS
Some news end of Jan 2012 about a meeting.
http://www.helenos.org/
Hurd
Difficult to scour news events from this site.
http://www.gnu.org/software/hurd/hurd.html
Inferno
Don’t have a news section but a mailing list.
Don’t know how active that is.
http://www.vitanuova.com/news/newsgroup.html
Kolibri OS
Last release in 2009
http://wiki.kolibrios.org/wiki/Main_Page
http://www.l4hq.org/
Has current news.
“The L4 community had some news last month.
NICTA and Open Kernel Labs have announced a joint public release of the formally-verified seL4 microkernel. The release, for non-commercial and evaluation use, contains seL4 kernel binaries for ARM and x86, documentation, user-level example code, x86 Linux on top of seL4, and the formal specification of the kernel for the ARM platform.”
MenuetOS
Site not very news friendly.
Most .x1 “Bugfixes and improvements”
Mailing list would probably be best here to scour for news.
http://www.menuetos.net/
MonaOS
Last news not to long ago. Aug 5
http://www.monaos.org/
Plan9
Mailing list has news.
Feb saw some news release me thinks.
http://plan9.bell-labs.com/plan9/
Puredarwin
Not quite sure. The news does not list dates.
The attachement to the last news item is dated to 2009 though.
http://www.puredarwin.org/developers/qemu
StormOS
Some news Oct 2011 about a collaboration with Debian.
http://www.stormos.org/
Capros
Not easy to glance news from the site.
http://www.capros.org/
SkyOS
Last news seem to be 2008
Breadbox
Last news 2009
http://www.breadbox.com/news.asp
Visopsys
Has a news section and has current news events.
http://visopsys.org/about/news.php
Unununium
Some news but not very meaty.
http://unununium.org/
Maybe these project has current developments.
It is just not that easy to find.
And also there is other projects not listed here.
Edited 2012-02-08 14:46 UTC
And I believe L4 doesn’t really belong on a list of niche OS projects.
http://en.wikipedia.org/wiki/L4_microkernel_family
It seems to run radio stack and/or isolate it from the “user OS” – so for example many (most?) Android phones might be also running L4, by the virtue of using Qualcomm chipsets.
Edited 2012-02-08 15:21 UTC
Didn’t you noticed “Nightly builds” link? Last version was build at 6 Feb 2012. Few days ago was added driver for Intel videocards. Also, Kolibri has tons of new ports.
Did not notice it.
Nice to hear about the progress on Kolibri. You must submit pal.
In fact, we’re afraid that new distributive can’t be stable yet, so we don’t publish new version yet. Previous stable version was 0.7.7.0 (svn revision ~1350). Today’s nightly build (svn revision ~2350) should have number at least 0.8.5.0 because was added very important things (ext2/3 support, Intel video support, HD audio support, new mutexes, new disk subsystem, ACPI, new almost fully compliant POSIX compatiblity RTL, PE DLLs, tons of new ports). But all that features aren’t well-tested. We are shorthanded and need new developers and testers, but it’s hard to find them.
That is very nice features added.
If i’m not mistaken Kolibri is mostly written in machine language.
Does this make finding contributors even more difficult?
Kernel is written entirely in assembly, and 80% of software too. But API is very easy, and assembly coding for Kolibri is much easier than C coding with WinAPI. Also, we have ports of C, Lua, Python and Basic, and cross-compilers for C++ and FreePascal. I did few ports by myself, and I think that few new ports of non-assembly software (Java, AbiWord and WebKit, for example) would be nice.
Would like to pick your brain some more:-)
I saw Menuet has closed source it’s 64bit version.
Would like to know your thoughts on the challenges and advantages of a possible 64bit version of Kolibri.
Yes, Menuet64 is closed-source and non-free for commercial use. Also, mp3 player for Menuet64 isn’t free too. Menuet32 is died. That’s why we forked off. Now Kolibri kernel has not more than 20% Menuet code.
Our developers have thoughts about SMP and 64 bits. One of possibilities: launch one core in 32 bit mode, and second (and may be others) in 64 bit mode. So, with such technology you can launch both 32 and 64 bit applications without any problems and compatibility layers. 32 bit mode is enough for many desktop applications, and 64 bit mode is for apps using math library, for example.
Thanks for sharing this all. Is very interesting.
In conclusion of our discussion:
http://builds.kolibrios.org/ – nightly builds (sorry, only bare kernel and 60-70 additional programs, such as mp3 player, text, table, graphic editors, simple graphic viewer, z80 emulator, but without additional drivers, videoplayer and other ports)
http://websvn.kolibrios.org/ – SVN repository can be viewed online
http://board.kolibrios.org/ – most active resource, where you can find all KolibriOS developers
There are at least 6 years since I discovered this project. I hoped that they’ll close the gap but they were always far behind Windows. Windows is a moving target and even if they’ll come up with a fully functional Windows XP alternative, by that time Windows 9 or maybe 10 will be released and will have very little in common with NT 5.1/ Windows XP.
I don’t know why this project makes very little progress. Look at Haiku and BeOs. Haiku not only reached BeOs compatibility but even surpassed old BeOs in terms of quality.
More love? (towards lost darling, BeOS)
Whereas with ReactOS… not only the original is around and doing fine, I guess most of the potential “in love” devs aren’t that much into OSS and/or would rather aim to work at Microsoft.
You cannot compare having a bunch of passionated, skilled but volunteer developers that dedicate some of their spare time reverse-engineering Windows and writing ReactOS with having hundredths of full-time employees that have all the internal APIs well documented.
I think that ReactOS should look at the amazing system they have implemented, starting to make it stable with the features it has right now and start to grow in its own direction; South America, Africa and all the developing countries would be nice markets for OSes like this.
Getting the level of functionality that Windows has (that reinvents itself every 4 or 5 years) is just an unreachable utopia.
Edited 2012-02-09 00:08 UTC
If anything, this is one of the reasons why ReactOS is desirable. There’s a huge amount of software that runs on Windows XP, and a lot of people (particularly in the business world) are dependent on the OS behaving/operating in a particular way. Newer versions of Windows are changing a lot, and this has caused problems in the past (and will continue to do so) so it would make sense to have an open OS that will be available/supported after MS finally kills off XP.
At the same time, you could argue that by the time ReactOS is usable, WINE will probably cover most bases, and that people may have moved on from XP. Even then, it will still be relevant in the way FreeDOS is.
Caveat emptor: I have never attempted to design an operating system.
But, I’ve followed ReactOS’s changelogs assiduously since God knows when, and it always strikes me that they’re constantly making infrastructural improvements to their build system, CVS, testing framework, etc. As an OS dev layman, these things strike me as things that should’ve been done from the get go, and each time it feels like, “Aha! Now the floodgates will open and they’ll really be able to start progressing by leaps and bounds!” …And time and time again, I’ve been disappointed.
I mean, a kernelspace testing infrastructure as part of the last release–isn’t it about time for such a major project?
Do my sentiments make any sense? I don’t mean to criticize, I’m genuinely ignorant and curious.
At any rate, kudos to the team! I’m a big fan and wish them all the best.
Well, if they are continuously changing it, then that could hinder developers as once developers get comfortable enough to try to participate, it changes. So it bites both ways.
That said, as you build out, you learn of new things you need to do; new capabilities you need; etc. I haven’t followed them closely, so I don’t know how much of is change for change sake or how much of it is to meet real needs, etc.
Shame not enough good people supported it. Should be on version 10 by now. Using wine instead brings more interaction so this was the proper way to try to use windows drivers and programs I think. Wine was a cure for those who ran other OS’s mainly and needed a small subset of windows applications.
This will bring more attention to the Windows users around the world. Might as well make a server edition of ReactOS in the near future.
I’m sure it will give just as fruitful results as Syllable Server…
(and while you’re quite possibly technically correct in the first sentence, the “more attention” increases would likely be measured in per mils, at best)
The need for ReactOS is clear. There is some software that will never run on WINE. Also, there is some hardware for which only Windows drivers are available.
You could just buy Windows but ReactOS brings all the benefits of Open Source. Look at the response to MS introducing the Metro UI and removing the start button. If I was a business that did not want to retrain all my employees, I might choose to stay on the old interface if I had more control over the code.
You probably wouldn’t be a very good business (boss?), if micromanaging about niche tech toys like that instead of focusing on the, well, business… (parts of which can be done, at worst, in virtualised XP environment)
Edited 2012-02-14 23:52 UTC
Just as Linux now evolves more quickly than the old UNIXs, ReactOS may someday outrun MS. That may sound fanciful but I remember the Solaris (and other) guys saying the same thing.
The problem ReactOS faces is that it may be sometime before it sees any commercial backing. Linux really benefited from commercial support, first home-grown and then from more established players.
Still, I would hardly bet against ReactOS at this point even though it’s day in the sun might be a way off.
ReactOS is taking a long time but I think it may evolve faster than history suggests.
First, it seems much better managed lately. It used to seem quite hostile to users and potential new devs. After many years of failing they finally were accepted into the Goggle summer of code. They have even been making a little money lately here and there.
Second, people may not realize it but they lost a lot of time doing an audit. They were worried at one point that somebody may have contributed MS code so they spent almost a year vetting the code. They lost not only time but some devs in the process. This is now behind them.
Third, they seem open to cooperation with other projects. For example they just got USB support from adapting code from Haiku and I know they contributed patches back. They also improved the regular networking code using code from elsewhere I believe. Of course, they also codevelop a lot of code with WINE and they have project to enable using more of the WINE code as is.
Finally, I think they will attract more users and devs. ReactOS is on the verge of becoming much more useful. USB support, wireless, sound, etc make running it on real hardware a possibility. Also real software like Firefox, OpenOffice, and .NET are beginning to work.
There is real work to do but we may be pleasantly surprised by what apps start to work over the next few years. The guts of Windows have not been changing as much as some people think and many apps are written to work on versions of Windows that are ten years old.
One final note about the constant rejigging of build systems. I know one of the big pushes has been to get ReactOS to build with both GCC and Microsoft’s compiler. That effort largely complete and could also accelerate development from here.
Great points.
I just tried installing Photomatix in Wine to no avail on Fedora 16.
I got a giant red X across the menu bar.
I couldn’t re-size or move the MDI windows.
Unusable.
For my photos I use rawtherapee and GIMP. Both are excellent applications. But for HDR nothing comes close to what is available on Windows or Mac. I don’t want to buy a Mac, build a hackintosh, or even get it running under VirtualBox because I’d still have to pay Apple for the OS. Looks like ReactOS is my last resort before having to actually buy an OS.
I hope Photomatix works well under ReactOS on VirtualBox and that I can easily move files between there and my host.
It seems to me that in 2000 – 2008 there were much more interesting os-es being in active development and even interesting and exciting features added to old OS-es.
It’s just me that finds the current OS landscape grey and boring?
I’am long follower of alternative OS’s. KolibriOS is defenitely alive, though don’t have releases for 2+ years, since we wait for completion of USB driver (in asm…) before major release. I admit that based on svn activity, ReactOS is more active then Kolibri and Haiku, but activity can be diffirent – copying of Windows API and creating OS from scratch differs a lot.
Passion in alternative oses is a complex thing.
I followed in the past EVERY single news about a new, fabulous alternative operating system and dreamed about a better IT world only to escape the boring and frustration of using Windows or a “then horrible and unfriendly” os (Linux anyone?).
Now things are changed, we are all olders and mature.
I work with Windows, I evangelize Linux and 1 Month ago I bought a Mac to develop.
The only interesting projects out now, IMHO, are Haiku and ReactOS (in the sake of ol’ better world dreamings) and Visopsys for diagnostics on malfunctioning systems.
As always, younger rebellious people, inevitably seeing “wrong” with the world, got new toys back then (PCs); fluctuations of a new unknown & exciting field.
…but just like with, say, model airplanes it calmed down after a while, after enthusiasm died out, project goals mostly revealed themselves as overenthusiastic.
Oh well, life goes on & now young devs seem to be on the mobile bandwagon, we’ll see what this one will give.
Edited 2012-02-15 00:02 UTC