99 man effort.

[Haos]

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.

Does the lack of proper proof defend your case and stands against mine? I dont think so.

Nute asked me to name one application not working in 0.3.0 and working in current version. I named two, popping right from my memory. There are loads more. Iirc, 0.3.0 had no MESA, thus no opengl app would work. Those have chance to work now.

I have been with this project since 0.3.0/0.3.1 transition and believe me, i seen many wondrous things happen. Its true that most part of changes do not appear anywhere for casual user, like having an app working. Most important changes from 0.3.0:
- stable kernel (try running 0.3.1 and doing anything with OS itself);
- UNIATA;
- new loader, 2003 compatible;
- sound support (working with real hardware)
- virtualbox support (with 3d acceleration)
- new network stack
- initial windbg support
- PSEH2;

This is again from top of my head, things that are kinda important to me. Its kinda hard not to notice those.

[Lone_Rifle]

I just realised that nute was reading wikipedia. I have been editing that article in the hopes that it would one day become a Good Article, and then a Featured Article, so I need to shed some light as to how I arrived at the 99 people:

For an article to be considered Good or a Feature, Wikipedia does not allow self-published sources of information, that is, information from what is being described in the wikipedia article itself, unless the entity happens to be a company of considerable repute, who risks catastrophic consequences if their information is found to be inaccurate or misleading. As such, we cannot use information sources originating from ReactOS itself.

The only way that I can thus comment on the number of people that work on ReactOS is to get the information from a secondary source, that is, information that does not come from the entity being discussed in the Wikipedia article. In this case, I used a feed from Ohloh, that tracks SVN commits from developers. This is found here:
http://www.ohloh.net/p/reactos/contributors

The drawback of using Ohloh is that they record all commits, so they also record people who have committed in the past, but no longer commit at present. This is why the ReactOS article reads as follows: "While the ReactOS entry in Ohloh records 99 contributorsr both past and present..." : because I want to emphasize that the source of information records historical information, not current information.

(Before anybody says anything, the Ohloh source was at 99 contributors when I accessed it. This is reflected in the Wikipedia entry as follows: "ReactOS Contributors". Ohloh. http://www.ohloh.net/p/reactos/contributors. Retrieved 2009-12-16.)

Ged is correct in pointing out that the developer numbers listed for Windows 7 are for Windows 7 alone: The reason this was still included in the ReactOS wikipedia article is because it's the only accurate source of information I could find and reference that lists headcount.

As a side note, I remember that the topic is about headcount and driving the development effort. Could somebody tell me again why we are talking about application compatibility?

[livestrong2109]

I wish you the best of luck on the publication, I have been cleaning up the wiki off and on for a few months now and over all have seen much progress from other users as well. It does seem much more professional, though we are in great need of citation as you have stated. I'll try and back track and see if I cant find some basis for some of the article.

Keep it up your work is valued and will being some great P.R.

[Ged]

The drawback of using Ohloh is that they record all commits, so they also record people who have committed in the past, but no longer commit at present. This is why the ReactOS article reads as follows: "While the ReactOS entry in Ohloh records 99 contributorsr both past and present..." : because I want to emphasize that the source of information records historical information, not current information.

Why would you add something to wikipedia that is completley misleading?
That '99' figure is the number of comitters from the start of reactos, but you make it sound like that's how many we have working on it now.
No wonder nute got it wrong. This needs to be changed.

[Lone_Rifle]

Hello Ged,

I believe that the manner I have phrased the sentence clearly indicates that the information does not reflect current numbers at all, and that nute has not read the article properly. As per what you have quoted me saying, I note that the figure is the cumulative number of people who have submitted commits to SVN, and thus reflects the number of people who both presently and historically (ie, the "both past and present phrase" that I used) have worked on the project, and not the present number of developers.

I am unable to find any third-party information that documents the number of people who are presently working on ReactOS. As mentioned already, due to constraints enforced by Wikipedia, I cannot use self-published information, see [1], and hence I am forced to use Ohloh to comment on developer numbers.

[1] http://en.wikipedia.org/wiki/Wikipedia: ... d_paper.29

[Ged]

I believe that the manner I have phrased the sentence clearly indicates that the information does not reflect current numbers at all, and that nute has not read the article properly

I disagree, I don't think it's made compeltley obvious, and the fact that the real number of active devs is about 10 makes it even more misleading.

I am unable to find any third-party information that documents the number of people who are presently working on ReactOS. As mentioned already, due to constraints enforced by Wikipedia, I cannot use self-published information, see [1], and hence I am forced to use Ohloh to comment on developer numbers.

[1] http://en.wikipedia.org/wiki/Wikipedia: ... d_paper.29

I'm not sure what your obsession with wikipedia guidelines is.
According to this quote taken from the link you provided

self-published material may, in some circumstances, be acceptable when produced by an established expert on the topic of the article whose work in the relevant field has previously been published by reliable third-party publications

we are allowed to publish our own information.

Why don't you just forget about the wiki rules and publish some correct information about the project. If it's questioned by the wiki maintainers then quote the above link, or remove it. If it isn't (which I'm sure it won't be) then where is the problem?

[Lone_Rifle]

I disagree, I don't think it's made compeltley obvious, and the fact that the real number of active devs is about 10 makes it even more misleading.

Duly noted, will try to rework the sentence to make it clearer. Please follow-up on the Wikipedia talk page.

A number of reviewers have criticised the ReactOS wikipedia article for the inclusion of self-published sources. This has happened twice already, both during Good Article reviews. We were able to pass the second Good Article review after I removed any references to our website where possible. It is hoped that by working towards Good Article, and then Featured Article, we can one day get some manner of publicity and notability; bear in mind that Featured Articles will occasionally be used to front Wikipedia's Main Page. The article is only allowed to reference self-published information under the circumstances that you have quoted.

Besides which, if we want to gain favour from Wikipedia in the form of publicity and notability, we should play by their rules, no?

[Ged]

I think the importance of getting good and correct information into wikipedia is far more important than getting a 'good article' status.
I've been using wikipedia since it's conception and I had no idea there was even such a thing.

Anyway, you're the guy maintaining it so it's up to you what you think is best

[Lone_Rifle]

Appreciate your acknowledgement of my autonomy. It occurred to me that when fireball wrote his call to arms for Arwinss, he could have noted the number of current developers working on ROS. That way OSNews would have made note of it, and the information can be referenced...

[hto]

I agree with Lone_Rifle about following Wikipedia rules.

Phrases like "active developers" or "current developers" have little sense. Can somebody say what does it mean, then it can be calculated by a mechanical way, and some third party can be asked to verify and confirm it.

[zefklop]

What about counting people that made commits between the last two releases and adding the new ones since then? That would make sense, and the information is easily available from changelog.

[hto]

What about counting people that made commits between the last two releases and adding the new ones since then?

Ok, counting commits in the svn://svn.reactos.org/reactos/ repository from (but not including) revision 41757 (release 0.3.10) to revision 44687 (release 0.3.11) gives 31 people:

• fireball (385)
• tkreuzer (349)
• janderwald (335)
• sginsberg (248)
• cgutman (174)
• gschneider (168)
• cwittich (164)
• dgorbachev (131)
• jimtabor (123)
• dreimer (108)
• dchapyshev (102)
• ros-arm-bringup (94)
• sserapion (75)
• hpoussin (66)
• arty (55)
• hyperion (47)
• spetreolle (38)
• mkupfer (38)
• khornicek (36)
• gedmurphy (35)
• ablackmann (31)
• pschweitzer (27)
• mjmartin (26)
• cfinck (25)
• lsuggs (15)
• gadamopoulos (9)
• jmorlan (8)
• bfreisen (6)
• ahill (6)
• gbrunmar (4)
• weiden (2)

And since then (to r45185):

• tkreuzer (93)
• cwittich (48)
• jimtabor (42)
• fireball (40)
• ekohl (36)
• sir_richard (35)
• ros-arm-bringup (27)
• sserapion (25)
• gedmurphy (19)
• cgutman (18)
• gschneider (17)
• dgorbachev (17)
• janderwald (14)
• dreimer (13)
• spetreolle (10)
• arty (9)
• mjmartin (8)
• mkupfer (7)
• martinf (6)
• khornicek (5)
• dchapyshev (2)
• cfinck (2)
• bfreisen (2)
• sedwards (1)
• hpoussin (1)
• gadamopoulos (1)

"New" people are ekohl, martinf, sedwards and a certain mysterious sir_richard.

So, the number of active developers (by zefklop's definition) is 35.

[vicmarcal]

Plus a lot of contributors

[jorl17]

If you need a source, then point wikipedia to this thread, all I can say.

[nute]

As far as hopeless goes, a project that chronically does not have enough developers may be hopeless.
A few applications working, and may I remind people that the commercial version of Diablo II does
not work on ReactOS, does not seem all that impressive for 10 releases of ReactOS. Is it possible
that ReactOS had more developers at one point in the past than the 5-10 active ones it has
now? If ReactOS can't be completed in a reasonable amount of time, will computer hardware
and software change enough in the mean time that noone wants it?

It is difficult to clone a closed source operating system, admit people that that is what this
project is about after all. The decisions that have to be made are hardly your own, you aren't
designing the specification after all. You don't get to decide what data structures their will
be. Regressions are highly likely because you are trying to run compiled closed source software
that was designed for the proprietary closed source system.

Already, most people are running 64 bit WIndows 7/Vista where ReactOS, if I'm not mistaken, is a 32
bit system. Maybe this isn't a big deal, in theory a 64 bit compiler can produce a 64 bit version of
ReactOS. ReactOS could use 99 or more active developers.