Linked by Eugenia Loli on Sun 10th Aug 2003 12:26 UTC, submitted by Gsurface
ReactOS For those that have no idea what ReactOS is, Flexbeta took some time and explained it, covered it's current status and write about where they plan to go. See the full review here.
Order by: Score:
Great idea but progress is too slow.
by walterbyrd on Sun 10th Aug 2003 13:11 UTC

By their own admission, ReactOS is years behind NT 4.0, a product that MS released in 1996. Considering the pace of their progress, ReactOS might have something almost usable by 2006. Will the ReactOS idea even be relevant by then?

It too bad. ReactOS could have been a much greater threat to Msft that Linux or MacOS ever were.

RE: Great idea but progress is too slow.
by jtd on Sun 10th Aug 2003 13:36 UTC

Um, you know that Wine dose not implement all Windows 2000 and XP APIs and runs lots of windows programs, do you? The idea is to implement a substantial part of the API before going on to implement less used ones. That is what ReactOS is doing now.

Reviews....
by bobbob on Sun 10th Aug 2003 14:46 UTC

Please can OSNEWS stop calling an overview of an installation a review?

Cheers,
bobbob

Alternatives are a good thing
by abraxas on Sun 10th Aug 2003 14:54 UTC

This sounds like an interesting project. The real issue is whether or not it will attract enough attention to become a viable alternative. Unfortunately it seems that by the time something usable is created, Longhorn will have been released and ReactOS will have to play catch up again. If Longhorn happens to break enough programs or drivers though, ReactOS could fill in nicely for people with legacy hardware and software. The future of any OS is murky at best at this point in time but I wish the developers of ReactOS the best of luck. Choice is a good thing.

re: bobbob
by krammit on Sun 10th Aug 2003 16:24 UTC

95% of the time, I'd agree with you. In this case though, it seems to be an accurate review. It doesn't seem ReactOS is capable of anything else yet. Calling it an introduction to ReactOS might have been a better billing...

RE : Alternatives are a good thing
by Aditya on Sun 10th Aug 2003 18:50 UTC

I totally agree with you abraxas sure it is much more interesting OS project a nice alternative to windows when it will finish
& good thing is that they are not going to build it on top of unix like everybody. however it is gonna need a lot of support of open source community in order to be alternative to Longhorn however this seems difficult as 95% of open source developers claim that windows suck (which is not at all true) so they might not contribute to this project. however it is the one to keep an eye on!

cloning Windows: API or kernel
by Ulrich Hobelmann on Sun 10th Aug 2003 21:56 UTC

Ok. People want to run Windows apps, certainly.
But there's lots of people who want to run an open-sources kernel that they can monitor security-wise and modify to their needs. Thats why we have Linux, BSD and OS X .
Of course there's two approaches to implement Windows emulation on another OS: emulate the whole API, or emulate the kernel.
The WINE people try to implement the whole Windows API, so they give you a completely free implementation.
The NetBSD people emulate Linux and other Unices with simple kernel emulation, thus they don't have to rewrite the whole API.
Hmm. I guess who writes the most stable implementation of the MS APIS..., MS?

What I actually want to say is...
WHY spent YEARS writing a buggy API replacement, when you can just emulate the kernel to allow Windows programs to peacefully run on your Unix or whatever system?
I have binary Linux libraries (SuSE's I think) installed on my NetBSD box, so I can run binary Linux Java JDK.
I wouldn't care if I had to install files from a legal Windows XP distribution to allow apps to run.
And certainly it would be 80% faster than WINE.
Why not just take the short path and then replace the libraries one by one (like OpenBeOS do)?
For one the city of Munich wouldn't have had to buy 1000s of licenses of VMWare to run Windows apps that way (yeah, maybe some simple Windows licenses would have been cheaper for them than running Windows in a Linux emu ;) ).

re:Alternatives are a good thing
by DavidGentle on Sun 10th Aug 2003 23:18 UTC

But surely if longhorn breaks a lot of programs people will want an alternative that doesn't. So they might choose reactos just to keep running what they already have on NT 4.0.

emulation
by xmp on Mon 11th Aug 2003 00:23 UTC

Well emulating Windows at kernel level would be a bit different than linux on bsd. We have no access to the kernel source and that's the crux of the problem. Correct me if I'm wrong.

As far as quality ... wine is still in alpha, ReactOS is more primitive than even that, while Crossover Office and WineX are useful to some extent. Meanwhile, Cygwin and various other Unix on Windows emulators seem quite advanced in comparison.

xmp

Re: emulation
by walterbyrd on Mon 11th Aug 2003 03:36 UTC

>>Meanwhile, Cygwin and various other Unix on Windows emulators seem quite advanced in comparison. <<

Which would only make sense since Linux and BSD are open, and windows is closed. Also, msft will do everything they can to keep windows apps from running on anything other than windows.

"For one the city of Munich wouldn't have had to buy 1000s of licenses of VMWare to run Windows apps that way (yeah, maybe some simple Windows licenses would have been cheaper for them than running Windows in a Linux emu ;) )."

The legacy Windows apps are Windows 3.1 based. Windows XP licenses (or a useable ReactOS for that matter) will be of no use at all for those.

Windows kernel emulation ; VMWare
by Ulrich Hobelmann on Mon 11th Aug 2003 09:12 UTC

@Guido: Thanks, I didn't know that. But doesn't WINE quite well emulate 3.1? At least I find VMWare _very_ expensive.

@xmp: You're right in that implementing a Windows (kernel) emulation layer is waay harder than a Unix layer, since there's no open reference implementation, and since Windows is so different from Unix (the Linux layer on NetBSD is fairly small).
But almost no programmer should need to look at implementation, but rather at interface specifications (i.e. Javadoc, C header files and the obligatory manual for production software). A lot of open-source software (sadly) is badly or not documented, that's why you need the source there. But for well written interfaces it should be possible.
Still I suppose MS does its best to hide certain things, and even the OpenBeOS crew seems to have problems with some "undocumented features".
I wish there was a law that required people to publish open hardware and software _interfaces_ for the sake of interoperability, especially now that more and more countries make it illegal to reverse-engineer and "break protection mechanisms".

RE: re:Alternatives are a good thing
by zima on Mon 11th Aug 2003 10:24 UTC

But wouldn't they just stay with NT 4.0 instead? I really think so...
Some will say that NT has no support...but which support will be worse, ever? The one of NT 4.0 (think about how many people is familiar etc.) or ReactOS?

That's the thing that struck me first. It's a nice bailout for NT users.

And that's also another nice thing about the Free/Open Source Software (FOSS) community - if you don't like X, then Y is there to help. If you think Linux sucks, no one's going to stop you from using OpenBSD or FreeBSD.

And if there is someone out there who knows a lot more about Windows NT than me, and who prefers FOSS - here's your opportunity to shine.

> Also, msft will do everything they can to keep windows
> apps from running on anything other than windows.

There's WINE, there's WineX, there's CrossOver.
MS did nothing against them. Stop trolling.

Re: KAMiKAZOW
by walterbyrd on Mon 11th Aug 2003 13:19 UTC

>>There's WINE, there's WineX, there's CrossOver.<<

All totally inadequate. Try to run a common application like ms-project on any of them.

>>MS did nothing against them. Stop trolling.<<

MS constantly changes the default files formats and like. Linux can never catch msft in this manner. Because if linux ever gets close, msft will simply change everything.

walterbyrd
by KAMiKAZOW on Mon 11th Aug 2003 13:48 UTC

> Try to run a common application like ms-project
> on any of them.
MS Project is no common application.
IE is a common application and IE works.


> MS constantly changes the default files formats and like.
> Linux can never catch msft in this manner. Because if
> linux ever gets close, msft will simply change everything.
MS doesn't change file formats, APIs, etc. very often.
If they change anything, they also break compatility with ther own applications. One of MS' keys to success was/is to be (mostly) compatible with their own products.
You're trolling. I won't any longer argue with you.

re:walterbyrd
by schmegglefurt on Mon 11th Aug 2003 15:10 UTC

>Which would only make sense since Linux and BSD are open, and
>windows is closed. Also, msft will do everything they can to keep
>windows apps from running on anything other than windows.

Which explains why they sell and maintain VirtualPC for Macintosh....or does it?

RE; re:walterbyrd
by dizz on Mon 11th Aug 2003 16:09 UTC

>Which explains why they sell and maintain VirtualPC for >Macintosh....or does it?

well isnt Virtual PC just what the name is implying
a virtual pc. witch means that all the windows apps that you are running under vpc is running under windows.
o fine vpc runs under OSX but that dosent matter you still need a valid windows licens.

anything else would be to say hey linux can run windows apps just fine cus we got wmvare

Re: KAMiKAZOW
by walterbyrd on Mon 11th Aug 2003 17:01 UTC

Are you really unable to see my point, or are you just being thick-headed and argumenitive?

Forget Wine, WineX, and CrossOver office. They are *far* too limited. They run a very small subset of windows apps. And crossover office is expensive on top of it.

ms-project is not that obscure. There are thousands upon thousands upon thosands of windows apps that won't run in linux with any of those products. And a lot of people need those apps. This is what so many desktop linux advocates don't get. The advocates figure if they can get email and play mp3s that's all anybody will ever need. The real world of IT is very different.

I wish the situation were different. I happen to like Linux. I think the msft way of doing business is contemptable. But I will not stick my head in the sand and deny reality.

Again. ReactOS is trying for binary compatibility with NT 4.0 - a product from 1996. And ReactOS is years behind even that. I am sorry to report that the ReactOS project is moving far slower than msft. By the time ReactOS has anything useful, msft will have moved to a very different model.

From this review..
by mario on Mon 11th Aug 2003 19:15 UTC

..I conclude that ReactOS blows. What an embarrassing thing to show, after all this development work.

Or am I being unjust? Maybe these guys really worked very hard?

Re: From this review..
by walterbyrd on Mon 11th Aug 2003 20:43 UTC

>>Or am I being unjust? Maybe these guys really worked very hard?<<

You are being harsh, but not unjust. Yes, they worked hard. But they going up against a super-rich and super-aggressive company.

Msft has every advantage. These guys have to find a way to do what msft has already done - they will always be trailing msft.


MS
by JSplice on Tue 12th Aug 2003 19:05 UTC

And MS will always be trailing behind Apple's new OS architecture, which is the best.

JSplice (IP: ---.gianteagle.com) - Posted on 2003-08-12 19:05:11
by Kingston on Wed 13th Aug 2003 05:43 UTC

"And MS will always be trailing behind Apple's new OS architecture, which is the best."

Though I do not fully agree with your statement, you are correct that in some ways Microsoft is falling behind. Quartz Extreme being a good example, and the updated OpenStep API being another.