99 man effort.

Here you can discuss ReactOS related topics.

Moderator: Moderator Team

nute
Posts: 251
Joined: Tue Aug 29, 2006 1:30 am

99 man effort.

Post by nute »

I hear that ReactOS is approximately a 99 man effort with no current corporate backing. Compare this to the 25 teams averaging 40 developers per team that developed Windows 7.

I went and took a fresh look at alternative operating systems. Are the Windows APIs/ABIs the most well thought out and efficient APIs/ABIs in existence? Syllable, Visopsys, Menuetos, Minix 3, etcetera are all alternative operating systems that may offer similar functionality eventually, but not
binary compatibility so much with Windows software and drivers.

One thing I don't understand about ReactOS is why undocumented Windows function calls are so important?

Is it possibly to increase the number of ReactOS developers by a factor of 10? What is keeping the number of developers working on ReactOS down?

Ged
Developer
Posts: 925
Joined: Thu Sep 29, 2005 3:00 pm
Location: UK

Re: 99 man effort.

Post by Ged »

nute wrote:I hear that ReactOS is approximately a 99 man effort with no current corporate backing. Compare this to the 25 teams averaging 40 developers per team that developed Windows 7.
Current active reactos devs are about 5-10 part time. If you followed reactos like everyone else, then you would know this

The 40 man teams you speak of at Microsoft are feature teams. These are specalist teams working to improve existing, stable areas, they aren't writing an operating system from scratch.

Windows actually incorporates thousands of full time developers over the past 30 years.
Windows 7 wasn't written from scratch, it's built on all the previous versions of windows.

If you're trying to compare reactos development with windows, then you need to rethink.

RaptorEmperor
Posts: 247
Joined: Mon Jun 26, 2006 12:04 pm
Location: Baltimore, Maryland, United States

Re: 99 man effort.

Post by RaptorEmperor »

Nute, if you want something to get excited about, look up Arwinss.

http://www.reactos.org/archives/public/ ... 12709.html

There's a PDF at the end detailing who, what, where, and when of Arwinss.

Arwinss is going to allow ReactOS to re-use a lot more Wine code than we already are using by mixing the old Win32 architecture with the Wine architecture. This will allow us to have a working Win32 and free up what developers we do have to work on other important issues, such as USB support. Wine has hundreds of developers, so taking advantage of their efforts will effectively add to our own manpower, since the Wine team will effectively be writing much of Win32 for us.

dark
Posts: 278
Joined: Wed Apr 06, 2005 9:40 pm

Re: 99 man effort.

Post by dark »

RaptorEmperor wrote:Nute, if you want something to get excited about, look up Arwinss.

http://www.reactos.org/archives/public/ ... 12709.html

There's a PDF at the end detailing who, what, where, and when of Arwinss.

Arwinss is going to allow ReactOS to re-use a lot more Wine code than we already are using by mixing the old Win32 architecture with the Wine architecture. This will allow us to have a working Win32 and free up what developers we do have to work on other important issues, such as USB support. Wine has hundreds of developers, so taking advantage of their efforts will effectively add to our own manpower, since the Wine team will effectively be writing much of Win32 for us.
I don't believe there's any official word on whether that will be used yet. Although I haven't noticed much difference since 0.2.7 so I'm okay with abandoning some compatibility to get a usable system for now.

User avatar
EmuandCo
Developer
Posts: 4439
Joined: Sun Nov 28, 2004 7:52 pm
Location: Germany, Bavaria, Steinfeld
Contact:

Re: 99 man effort.

Post by EmuandCo »

Sorry, but if you did not notice any changes since 0.2.7 then you did not try ros or you are blind.
ReactOS is still in alpha stage, meaning it is not feature-complete and is recommended only for evaluation and testing purposes.

dark
Posts: 278
Joined: Wed Apr 06, 2005 9:40 pm

Re: 99 man effort.

Post by dark »

EmuandCo wrote:Sorry, but if you did not notice any changes since 0.2.7 then you did not try ros or you are blind.
Yes there is the limited SATA, usb, sound, networking support, etc... but in terms of new applications working, not much at all. Many of them seem to regress for releases as well. Usability wise for most people, it seems pretty much the same. I suppose the setup APIs were implemented since then, and I think that's the most useful thing I can remember.

Haos
Test Team
Posts: 2954
Joined: Thu Mar 22, 2007 5:42 am
Contact:

Re: 99 man effort.

Post by Haos »

You are trully blind, or got bad luck, so the bunch of apps you are interested in had no progress. Still, bad luck.

nute
Posts: 251
Joined: Tue Aug 29, 2006 1:30 am

Re: 99 man effort.

Post by nute »

I will defend Dark. I've been trying every release since at least 0.3.0 and I can attest to the lack of visible improvement. Dark isn't so blind really. Name one application that works in ReactOS now that didn't work in 0.3.0? Yes Firefox 2 works better now than it did in the 0.3.0 days, but Firefox 3 still doesn't work. A 5-10 man/woman project with noone dedicated to it full time runs the risk of progressing slower than a 1 man project. More developers are needed, why aren't there enough? I bet ReactOS had more developers at some point in time than it does now. Berating Dark doesn't help the situation. Maybe 5-10 isn't the number of developers really considering that these developers only work on ReactOS when they feel like it which may not be very often at all. Linus' law states that all bugs are shallow if there are enough eyes on the code. Well, 5-10 is not enough for millions of lines of code. There needs to be a developer
drive and a fund drive in my opinion and it is beyond time to hire someone to make ReactOS their full time job. If this doesn't happen, I fear that ReactOS will be irrelevant for 5-10 more years at which point it is possible that noone will care about NT anymore.

User avatar
EmuandCo
Developer
Posts: 4439
Joined: Sun Nov 28, 2004 7:52 pm
Location: Germany, Bavaria, Steinfeld
Contact:

Re: 99 man effort.

Post by EmuandCo »

Nute... you are driving me nuts!

"I've been trying every release since at least 0.3.0 and I can attest to the lack of visible improvement."

Another example of a person who just tries Warcraft 2 and complains when it fails to work...

"Name one application that works in ReactOS now that didn't work in 0.3.0?"

SeaMonkey 2, DosBOX, Diabolo 2, abyss web server, MirandaIM, RosBE, VLC, utorrent etcetc

"A 5-10 man/woman project with noone dedicated to it full time runs the risk of progressing slower than a 1 man project."

Err, what?? Where is this lie from?

"More developers are needed, why aren't there enough?"

We can do this again, sure. It was just answered 3*10^8 times before. We do PR, but there are not many Devs who know of the Low Level stuff inside the NT Kernel. Go and find some!

"I bet ReactOS had more developers at some point in time than it does now."

Nope.

"Maybe 5-10 isn't the number of developers really considering that these developers only work on ReactOS when they feel like it which may not be very often at all."

Another thing we discussed even more often... I wont answer this again. Look through your other posts and answer it for yourself.

"There needs to be a developer drive and a fund drive in my opinion and it is beyond time to hire someone to make ReactOS their full time job."

Create one and pay us. There are no commercial supporters ye, so feel free to pay us instead.

"...at which point it is possible that noone will care about NT anymore."

Proof or did not happen
ReactOS is still in alpha stage, meaning it is not feature-complete and is recommended only for evaluation and testing purposes.

Aeneas
Posts: 470
Joined: Sat Oct 10, 2009 10:09 pm

Re: 99 man effort.

Post by Aeneas »

Please allow me to set out in the beginning that I do not wish to enter into any debates of any person here.

I read Aleksey's presentation. I am no developer so my ability to judge is limited. But I did see some interesting points in it. Apparently the core of the problem is whether to stick to the success of another project (Wine IS successful in terms of working applications), AND lose a lot of time invested, or - to stick with the current path, but possibly have to invest even more time, which, if without success, would create an even bigger waste of time.

I cannot judge what negative implications the increased adherence to "the Wine way" may have. But one thing is true - ReactOS is not, now purely from a user point of view - a user wishing to run his applications that is, working as well as Wine. Please do not misunderstand it as criticism, and even less as a lack of respect of the efforts so far, the results are in their own way remarkable. But most probably few people - if any - would like to use ReactOS for their everyday work as it is now. - But that would be precisely what ReactOS would need in order to "impress anybody".

So if there is some easy way to look and act like Windows for 90% of the software around, I would assume it is better to go it. Even though you work hard and very well, ressources are tight. Maybe you do not have to completely scrap your current Win32 subsystem, but just leave it for some more far-away day, when you have more developers, and possibly even paid ones - which you would only get if there is some tangible result (again, tangible from a purely user point of view).

RaptorEmperor
Posts: 247
Joined: Mon Jun 26, 2006 12:04 pm
Location: Baltimore, Maryland, United States

Re: 99 man effort.

Post by RaptorEmperor »

Many issues in ReactOS involve Win32, and by reusing more Wine code we can have a working Win32 up and running sooner. With Arwinss in place, more development can be spent getting the other parts of ReactOS running. When the rest of ReactOS is running like it should (USB, networking, driver support, whatever) then the devs can go back and make a more structurally accurate Win32 subsystem.

If Arwinss is put in place and it works as well as Fireball says it will, then I'm sure that it will get us some publicity. Almost all of the applications I test have issues with Win32, and I know from experience that some of these applications also run better in Wine than they do in ReactOS. If Nute knows anything about the programs he's running in ReactOS, the ones he says never ever work, then he'd probably find dependencies in Win32 that are being unmet, dependencies that Arwinss would fill. If using Wine's Win32 components more can get us up to Wine's level of software compatibility, then we'll finally be a viable platform. What Nute seems to be overlooking is that a working Arwinss means working programs. When more programs work in ReactOS, then people get interested in using it more. More people test ReactOS on their own, see the real potential of this project, and some of those people know how to program, so they decide they want in on the action. Then we can get more people than the ten or so that are working on ReactOS right now.

Implementing Arwinss doesn't necessarily mean that all the work on the current Win32 subsystem has to go to waste. The developers could have a gentleman's agreement to rework a proper Win32 once the rest of ReactOS is up and running. By that time, our team of developers will be larger, since the progress we'll have made by then will have made us a real force in the open source OS world.

A working Win32 isn't the only thing that Arwinss offers. It's a fresh new opportunity to show to the public (including potential new developers) what ReactOS can do.

User avatar
Pesho
Posts: 209
Joined: Mon Dec 10, 2007 1:16 pm

Re: 99 man effort.

Post by Pesho »

I'm all for Arwinss, especially if the spike in compatability will increase the public interest proportionately. I also liked the part where it says it will free up development resources for stuff such as USB and sound support.

Haos
Test Team
Posts: 2954
Joined: Thu Mar 22, 2007 5:42 am
Contact:

Re: 99 man effort.

Post by Haos »

Total Commander 7.04
Civilisation 2

Wanna list more?

dark
Posts: 278
Joined: Wed Apr 06, 2005 9:40 pm

Re: 99 man effort.

Post by dark »

Nute defended me, I've lost all will to argue.

Anyway, there are millions of applications for windows and we're talking about maybe 50 applications. The sample size for an accurate representation of progress would likely be at least 400 applications, and I don't think it's possible to get that much testing right now.

RaptorEmperor
Posts: 247
Joined: Mon Jun 26, 2006 12:04 pm
Location: Baltimore, Maryland, United States

Re: 99 man effort.

Post by RaptorEmperor »

dark wrote:Nute defended me, I've lost all will to argue.

Anyway, there are millions of applications for windows and we're talking about maybe 50 applications. The sample size for an accurate representation of progress would likely be at least 400 applications, and I don't think it's possible to get that much testing right now.
People get irritated with Nute on here because he has a habit of rambling on the forums about how hopeless ReactOS is, and shoots down every feasible option to fixing the deficiencies he sees with ReactOS. I figured that Nute, if anybody, should be excited about Arwinss, because it means we may hit beta sooner rather than later. Say something positive about the pace of ReactOS development and you will incur his wrath. I generally avoid arguing with him, tough.

Anyway, this isn't an argument, it's a friendly discussion. :D

According to Fireball's PDF file on Arwinss, Wine has been tested against at least 13495 different programs thus far. You can go to the Wine website itself and see all the programs they've got working. If people want software to test against ReactOS, they can go to the Wine compatibility database, see what's running on Wine right now, and see if ReactOS is running it as well as Wine. There are millions of applications for Windows, but Windows only has so many components to call upon. You're not testing to see if the software works as much as you are to see if the components the software calls to are working. Even though there are many thousands of programs for Windows, you may only need to test a few hundred to demonstrate that ReactOS or Wine are properly (or improperly) implementing certain components. Nowadays, not even all Windows programs work on Windows. Earlier today I tried installing an old game called Outpost for Windows 95 on Windows XP (with compatibility mode!) and the installer refused to run correctly, killing the process.

If all of us testers were to test at least two programs a day, I'm sure we could test 400 applications on ReactOS in a month. How many testers do we have now, anyway? If Arwinss runs as good as it sounds like it will, I'll have to do a compatibility testing binge for a few days to meet the challenge. ;)

The only reason I don't do more testing right now is that I'm busy trying to transfer from one college to another, so I'm busy, and when I do test software I keep running into the same bugs over and over again. I have my own little batch of benchmark programs and I'm looking forward to seeing what Arwinss can do to get them running. I did my first debug log to find out that crappy GDI support was causing SimCity 2000 to crash in ReactOS, and I tested it in Wine and it ran great. Since most of these recurring bugs I keep encountering are bugs in Win32, I'm really excited to see what the new Wine code can do for ReactOS.

Post Reply

Who is online

Users browsing this forum: Bing [Bot], Yandex [Bot] and 2 guests