The ReactOS project has seen some major progress during 2008. The project, which aims to create a Windows NT-compatible operating system, has published a ‘year-in-review‘ article concerning 2008, detailing the various area of work. It provides some interesting insights into the project’s development.
Last year was a busy one for ReactOS. After the kernel rewrite, which allowed them to stabilise the kernel, the gained the ability to direct attention towards elements slightly higher up the software stack, such as the graphics subsystem and filesystem support.
ReactOS’ graphics subsystem was old, full of hacks, and in dire need of attention. Several ReactOS developers have therefore spent time on what is essentially a rewrite of the Win32 subsystem. Lots of work needed to be done on this subsystem anyway; it had internal data and architecture structures that aren’t present in the Windows implementation, and those need to be removed. In addition, several interdependency issues in the Win32 subsystem need to be addressed.
Work is also being done on ReactX, the project’s DirectX replacement. ReactX is a very daunting and difficult effort, and as such, they put several interim solutions in place, such using Wine code to translate DirectX calls into OpenGL calls, which ReactOS already supports. They are also working on getting the actual DirectX runtime working on ReactOS, and they have already managed some basic hardware accelerated 2D rendering.
The Filesystem Runtime Library and the Common Cache implementation also need to be reworked. Currently, FAT32 is implemented as a sort of hack in ReactOS, since the two mentioned components aren’t very mature. CC functionality needs to be separated from the memory manager; the CC is not a separate component, but merely exposes functionality within the memory manager. By fixing up the memory manager, the team could start working on the CC elements. Work on CC is progressing, but large strides have already been made. The FsRtl wasn’t present during the earlier years of ReactOS, but it’s coming along nicely in the present.
Efforts to port ReactOS to the ARM anc x64 architectures are also under way, with both projects being able to initialise the first few parts of the operating system.
While I personally believe that ReactOS will always be chasing a moving target, with the target driving a Ferrari and the ReactOS team chasing it on a tricycle with two wheels, one can do nothing but admire the efforts and pogress the project is constantly making. Let’s hope 2009 will be as productive for them as 2008 has been.
Good luck to the team, their job is tremendous compared to the task.
True! Iirc, the ReactOS project has been around for over a decade, and the first few years must have been a real slog.
They’re making great progress now though. Good on them!
Yes but they still have a lot of work to do
One the one hand it is impressive, but on the other hand it is also sad
one can do nothing but admire the efforts and pogress
Absolutely. Its heroic. The true spirit of OSS.
Good luck to them! This is such a huge project, but has so much potential.
On the computers I’ve tried ReactOS on, it doesn’t even complete bootup. I tried recently too.
That’s been my experience too. I’ll try it again soon. Last time I posted that here, I think someone took me to task for not reporting it to them. This time I will. Thought I don’t think it will help. I use some unusual hardware ( not your more common dell, hp, asus stuff).
I tried one of the 2007 beta’s on my old spare PC (FIC PA-2002 with a Kingston Turbochip AMD K6-II 400MHz CPU and 112MB RAM—12MB of RAM was faulty) and it booted faultlessly. I would have wanted to try installing freeware apps like OpenOffice but didn’t.
In developing countries (like the Philippines) where buying 2nd-hand PCs and laptops is a viable option, installing a copy of a stable version of ReactOS and adding your office suite of choice is workable, as an alternative to Linux or XP.
They need not to chase a moving target as they will never catch up to MS. What they should focus on doing is getting to the point where you can do most of the things you can in Windows XP SP3 which is not a moving target.
If they can get it to that point then it would be pretty damn sweet!
But we will see where they go from here.
I totally agree on this one, XP SP3 functionality and stability should be the first end-target, while possibly incorporating some Linux technologies that could improve upon Windows and would have already if MS hadnt been such moneyhungry and uncreative gimps ..
The tech that could be incorporated is an in some areas improved GUI for example, while not breaking compatibility and such with windows.
You right, you right. I was getting a little ahead of myself. LOL!
They don’t really even need to aim that high. Windows 2000 seems like a much better target to me. If it can do what win2k does, then I’d be happy to use it. Once you’ve basically gotten there then you can look onwards, but win2k seems like a perfect goal.
Yes, but don’t forget this project started by chasing Win95…
And then it was NT4…
So, by saying that they don’t need to chase anything newer than XP or even W2K, you’ve already stated that their original target moved. The fact is, in 5-10 years time, everyone will be telling them they should be targetting Windows 7
The only way their target might stop moving and they can truly catch up is if that Ferrari breaks down, or runs out of gas in the middle of the desert (which could certainly happen, but I suspect the tow truck would be around shortly to remedy the situation).
The reason I said XP or Win 2K is because unlike Windows 95 (Which lasted less then 3 years) XP has been around for almost 10 years and is not going anywhere anytime soon. Windows 7 will not displace it because too many people like the look and feel of XP. So this is a good target to shoot for.
Ah, but there was also Server 2003, and XP64bit which *are* different (even though you’d probably like to believe they aren’t relevant). These had newer driver requirements for example.
Personally, I use XP64 at home, and work with Server 2003 quite a bit at work, and find them very pleasant. The user-interface is a very very minor part of the re-creation of Windows – so this is not what you should base your decision on. The re-creation of subsystems, APIs, and the driver model are much more important for the ReactOS project.
I too would welcome an XP as the target, but you have to remain realistic as well. The advantage, as you’re pointing out, is that XP has had a long run, with much success software and driver-wise, and this doesn’t seem to be dying yet. But in 5-10 years from now, I suspect many hardware manufacturers will cease to offer XP drivers, and the advantage that ReactOS is counting on – driver binary compatibility – will start to subside again.
In any case, I fully support this project, and want to see it succeed and flourish as well
Just a little nitpick– Win95 is not NT4, but rather MS-DOS 4. (The NT lines and MSDOS lines were split in development until they focused on the NT kernel more after ME and Win2K. –I read about this when I did more digging on the Win7 name.)
—http://www.microsoft.com/windows/WinHistoryProGraphic.mspx :: outdated, but still useful
Edited 2009-01-18 21:41 UTC
Yes I know this, but we were talking about the ReactOS target OS, not the history of Windows…
ReactOS was originally started after Windows 95 was release, in an attempt to re-create Win95, eventually they realized how futile that effort would be, and changed their target to NT4 at some point thereafter.
It wasn’t dos 4. It was windows 4. Dos 4 was from 1988. In fact, win 95 reported that the included DOS was version 7.0.
http://en.wikipedia.org/wiki/Timeline_of_x86_DOS_operating_systems
I actually think that’s what they’re trying to do at this point.
But, I think Windows 2000 is their immediate goal. First, achieve a status and functionality that is similar to Win2K (even the styles look like it), and then start implementing what is seen in WinXP and NT6.x onto it.
(Luckily, since they are trying to make something as reliable and stable as the NT5.x versions and MS is continually developing, the ROS team is able to look at what they’re doing and plan a bit ahead
)
I wish all the best to the ReactOS, it`s great idea and project. I think the milestones for them have to be two – ntfs and usb. Good luck!
Why make there own implementation of DirectX if they build an Windows compatible operating system?
Because a) DirectX is not OSS, and b) the DirectX license may well prohibit the use of it on non-Windows OSes.
JAL
Anyone porting the necessary bits of that to Linux kernal, and Wine?
They do collaborate with the Wine project, but there really shouldn’t be any kind of react code going into the linux kernel.
It is too bad that they do not have more people involved in the project.
If more people who wished that Windows was different would work on ReactOS rather than moving to Linux it would probably move a little faster.
I do not mean this in a “people shouldn’t use Linux” kind of way. Linux is great for a lot of things, and because ReactOS is not more mature I am going to try to shift some of the PCs where I work towards Linux. It is just that moving to Linux will cost me time spent retraining users on a new OS and on using new software. I can cheat, to a limited degree, by introducing software like OpenOffice now and the new OS later, but I will still need to introduce the users to each of these things, be it now or later.
I wish ReactOS was more mature. It would be nice to breath new life into older machines with an OS that users do not need to learn.
It would be nice to avoid Windows 7 without looking for software to replace “this” and without retraining people to do “that” instead of “this.”
The best example I have is Great Plains.
How do I move all of my users and data from Great Plains on Windows to some alternative on Linux… painlessly?
On the otherhand, if I could just run the Great Plains client on ReactOS it would solve SO MANY of the issues I will be facing in the future – no retraining, for one example, reduced costs for another.
I do not have the skills needed to help develop an OS so I am going to continue to keep an eye on ReactOS and hope. (That and not so little things like rewrite all of our VBA based Excel Spreadsheets as OpenOffice Calc Spreadsheets instead.)
I should probably mention, as well, that if, in the long run, Windows 7 turns out to be the best tool for the job I will likely deploy it. But because it is based on Vista and Vista does not work in our environment I do not see it as a viable upgrade, at this time. No hard feelings, I just do not see the point in trying to use a screwdriver when I require a wrench to do the job.
Edited 2009-01-20 16:22 UTC