Why I'm giving up on ReactOS for now

[techpro]

I have spent way too much time installing different builds of ReactOS to have a fair chance of seeing what it could do. Thanks to help in this forum I found out how to get files into a QEMU image, and by going back to an old release I managed to get Firefox to install. But it seems clear that ReactOS has a very long way to go before it can run the majority of software.

I was very interested in the idea of a free OS that would run the vast catalog of Windows applications. But the fact is, wine under Linux can run a much bigger range of Windows apps than ReactOS can. If someone simply wants a free OS to run Windows apps on, who cares whether it's a compatibility layer between the apps and Linux or a clone of Windows, especially if the surrounding GUI tools were made to work similar to those in Windows?

My motive in looking at ReactOS was to see whether it was stable enough to be used to create boot CDs that could run troubleshooting tools written to fix Windows problems that make Windows itself unbootable. At the moment, such tools are either Linux based, which means you have to know Linux and Linux dev tools to create them, or else they are Bart PE based, which means you can't create burn and run ISOs for people to use because they have to be built from the user's own Windows files.

A major problem with this was always going to be the lack of an NTFS driver for ReactOS. But this could perhaps have been overcome if it were the only obstacle.

The major problem making ReactOS unusable for me is that it seems to have a serious problem with programs created using Borland Delphi. I tried installing an application whose setup is created using Inno Setup, an install builder created in Delphi that is very popular with shareware developers (because it is very good, and free) and it crashed ReactOS totally. I tried running several programs of mine compiled using Delphi and I got various errors about writing to streams or icons being in an invalid format. I also tried some programs written in Visual DialogScript, a shareware interpreted language written in Delphi. A few ran in ROS 0.3.0 but they all threw various unhandled exception errors in the later versions.

I'm sure working on ReactOS is an incredibly interesting and useful learning experience, but when I look at the difference between the 0.3.0 version of a year ago and the current version I wonder just how long it will be before you have a version that's useful to mere users?

[preston]

I'm sure working on ReactOS is an incredibly interesting and useful learning experience, but when I look at the difference between the 0.3.0 version of a year ago and the current version I wonder just how long it will be before you have a version that's useful to mere users?

If you want to see the real progress you need to look under the hood. Check how many commits ReactOS has per day, how many bugs are fixed and how many new things are implemented. I can understand that in this stage of development for average user it must seem ReactOS is still the same crashing alpha OS as it was year ago. But when you consider all those major kernel and subsystems rewrites you will see ReactOS is evolving every day . I hope you will pay us a visit in the future to see how much effort developers put in this project.

[Reacter]

Give it a year, two, or three. We'll make landfall as 0.5 soon enough... See you then!

[elektrik]

I have spent way too much time installing different builds of ReactOS to have a fair chance of seeing what it could do. Thanks to help in this forum I found out how to get files into a QEMU image, and by going back to an old release I managed to get Firefox to install. But it seems clear that ReactOS has a very long way to go before it can run the majority of software.

To an extent, I understand your frustration at not being able to do much on the O.S., however without restating the obvious over and over again, this is *still* Alpha software. As a QA Engineer, I can't begin to explain how fine grain detail these programmers have to go down to to replicate an O.S. like this: This is probably an overly-exaggerated analogy, but picture trying to replicate a sand castle by duplicating each individual grain of sand-they'll get there eventually, but It'll take time, look ugly in the meantime and won't be very 'useful'...

As an everyday user, I won't use ReactOS for much the same reason as you, but for me, it isn't so much the *now* as much as in the future, and I do see much progress being made-to the point that, IMHO, it's a short matter of time before it becomes truly useful and elegant. I'm of course basing that on my own observation of the newsletters, release notes, etc. but I think we're going to see some really *big* process in the next 6-12 months....

Just my $.02....want change? =]

[techpro]

Yep, I do understand that this is Alpha software. But in Wikipedia I seem to recall it said that this project started around 1996. Must be the longest Alpha in computing history! It just makes me wonder whether it is possible to complete a stable Windows clone in a reasonable timespan without the same kind of resources Microsoft had. I mean, the work involved must be the same. Harder, possibly, because you have got to emulate the behavior of Windows without - in many cases - specific documentation of how each API is supposed to work, instead of being able to say "it works this way because that's how I wrote it."

I really would like to see ReactOS become a usable system some time soon. I like the Windows architecture and don't want to give up a lot of Windows programs but I am unhappy with the bloated behemoths Microsoft is forcing me to use. But when I surf this forum and see comments like "nobody is working on the networking code at the moment" or "nobody is working on the VGA drivers" it makes me think ReactOS faces the same problem Linux had 10 years ago, only worse, because I guess it has far fewer developers.

I feel a frustration with the whole free software movement. Linux could really have become a serious threat to Windows if developers had pooled resources and focussed on fixing the issues that put non-techies off instead of creating hundreds of me-too distros that are little different from one another apart from the name and the desktop theme they use and add nothing to the OS as a whole. If some of that wasted effort could be diverted into ReactOS development then the alternative OS scene would quickly become a whole lot more interesting.

[dreams]

For most devs ReactOS is a hobby project. They will only work on it in their free time, and only on the stuff that they think is the most interesting to them. That is why some components like network code isn't being worked on now.

Also, if you don't want a bloated Windows operating system, install Windows 2000.

[lhstech]

"Yep, I do understand that this is Alpha software. But in Wikipedia I seem to recall it said that this project started around 1996. Must be the longest Alpha in computing history!"

*************************************************

hmmm...i can't help but point this out...BUT...have you ever heard of HURD...lol...

http://www.gnu.org/software/hurd/hurd.html

HURD start date: sometime in 1990
current stable release: 0.2
project status: ACTIVE with participation in Google's summer of code for 2007

in their defense...

Linux is the big dawg...hardly anyone cares about HURD
GNU Mach: they are running on a microkernel @ v1.3 that they are combining with HURD

STILL...after all i know of computer history (which i'm sure someone will point out something else, of course...8P ) i think these guys win the "longest alpha in computer history" prize.

ESPECIALLY...since HURD is only a kernel where the ReactOS team is building an ENTIRE operating system from scratch.

So in my inestimable opinion...lol...that would make ReactOS the 2nd longest alpha in computer history...LOL...j/k

you guys are great...keep up the good work...

[GreatLord]

Hi
All release from 0.2.x and down is pre-alpha
and 0.3.x is the frist alpha seire we have

What I see in 0.3.x serire against 0.2.x already it is allot more stable and less bsod. But we surffer of usermode crash and wrong desgin gdi subsystem. so we are now in progress redesign the gdi subsystem.
and that mean also we writing a testcase for each syscall api for win32k
that mean less bug in the graphice part. This work will take most of this year, and maybe whole next year as well, we are only 3 people that
working on this on our spare time. sametime we are rewriting gdi subsystem we are adding new api and allot bugfix same times.

Then we got other programes doing masswork on like PNP, ntoskrnl
and other part as well. each week it is less 100-200commits for varus
bugs and new thing. I thunk u should checkout the 0.3.3 release when it
comes. it will be the most stable version in the histry I think. against all pervus version. I am not full happy with the stable in 0.3.3 it have not done any bsod for me, but allot user mode crashm when it crash, and also
less memory leaks, and handles of varus things.

[florian]

Yep, I do understand that this is Alpha software. But in Wikipedia I seem to recall it said that this project started around 1996. Must be the longest Alpha in computing history!

In the beginning of its development ReactOS was called FreeWin95 - a clone of Windows 95. The NT-clone ReactOS started 1998 and probably the old FreeWin95 was not very successful or useful for ReactOS. Therefore you might regard the effective time of development three years less.

At least that's what I guess.

[artguy10]

And software -- especially an OS -- ESPECIALLY when it is free -- ESPECIALLY when it has so few developers -- needs to spend a good deal of time in pre-alpha to get things somewhat working before an Alpha can be made. Then they can start working on the alpha (as quickly as a non-profit group of programmers can).

And at that, it is supposed to be in beta by what, 2008? (According to the roadmap page.) Actually, with the audit going, it will maybe make beta by 2009, and that roadmap should be updated. Version 0.4 was scheduled for 2007 (time running out!), and version 0.5 is scheduled for 07/08 (maybe end of 08? but not 07)

[oiaohm]

Number 1 Reactos is not something like Linux or BSD where you could pickup a complete documentation to how the internals are meant to work.

Ie Microsoft has kept a lot of data secret.

Having people reverse sections to find out how it works takes time. Please look at Wine (winehq.org) Its been running longer than reactos and its not complete yet in the user level. There problem is that documentation about windows userlevel api's from the MSDN are sometimes incomplete or missing. Yet those facts don't stop coders from using those functions.

The more interesting question is how much time has been spent to workout out how stuff is truly should work compared to coding.

Also Wine and Reactos have both suffered from the same error that costs a lot of time. Ok we don't know how this should do a hack. Bad news about that it comes back and bits it. Most of the hacks have been removed correct functions used these days in Reactos.

Comparing on time don't really work. Subtract amount of stuffups and documentation creation and Reactos is not doing too bad.

[Angelus]

Ok, let's see:
- thousands of people coding vs. around 40 souls
- people with dedicated time vs. people that uses their spare time
- people that earn for code vs. people that do it for free
- people that own all the sources vs. people that try to guess them
- ...

I don't know what other people think about what means a project like this but... I think that it's great the job that they are doing (take a look the commits -> http://cia.vc/stats/project/ReactOS ) emulating a complete close-source OS like the NT series.

If some people have no patience to see ReactOS as a stable product, it is their problem. But I don't think saying things like "I give up..." is a solution, it is their problem. Solutions are things like "I'm not a coder but I know about someone that could help this project".

This is an easy math equation: +People_Coding = +ReactOS_progress

[techpro]

I don't deny any of that. However, if you make ISOs available for ordinary user types to try, you must expect to get back ordinary user type comments from people who found they couldn't use it for anything.

I have no desire to detract from the work of the people who are working on it. However it actually seems as if progress has been backwards, since I found things that worked - or at least got further before they crashed - in 0.3.0 that did not work in the latest releases.

Not only that, but the Windows environment itself is not standing still. Someone made the point that the wine project is not complete. But it is inevitable that it never will be, because Microsoft will always keep moving the goalposts further away. The question is whether the ReactOS team is large enough to even keep them in sight.

It would be great if more developers were working on it, but there is nothing I, personally, can do about that. If I wanted to contribute, I can only use Delphi, which doesn't run on ReactOS, and even if it did, you won't accept Pascal code into the project.

[oiaohm]

0.2.x Reactos yes looked to work better. Had a lot of work around hacks in it that makes it imposable ever to load real windows drivers. So yes it worked but internals complete screwed up.

0.3.0 0.3.1 had these bad work around hacks removed. Creating masive unstable status. 0.3.3 should be more stable than 0.2.x ever was. Repairing the error of work around hacks has cost time. 0.3.0 only ever worked in emulators.

[GreatLord]

Hi

If we look what is left tio fix in trunk (devloping tree), it is readding back SSE support, fix npfs working again as it should.

They are only known regress from 0.2.x left in reactos devloper tree.
thinking of it is basic only known regress/left to fix betwin 0.2.x and 0.3.x
then it should people see huge changes.

But if we looking at 0.3.x
it got over 600 000 Bugsfix/new api/rewriting of allot api/translations or more. It is rug guess I doing for I known my self fixed over 100 000 bugs/implemet new thing in 0.3.x same goes for other devs like alex and other devs.

And if u reading a case stuiding outcast order by EU, U can found it reactos mailing list I think. They wrote under the time reactos exists it should have req 1000 programers that wroking 5 days each week 8hours each day. so on.

That mean our devloping team are wort 1000 programers that working 5 days each week 8 hours acoding the out cst. to say we are lazy are wrong

U need also understand that when ReactOS was frist created everyone thinking of getting stuff working. Not so much about desgin how it was done in WIndows NT. The Desgin was started be impement for real around 2004, we come to a point we can not continue devloper rectos
thanks to all wrong stuff, and we did not have testkit at all. Thday we are
working getting testkit for each api, and tetsing stuff hard. That mean
at end we will see less regress and the api works same as windows Nt 2003