OSDev.org

Let's not go grossly OT and bash newbies, we have other threads for that

In a group of n developers, how many of them do you get to agree on:

• C vs. C++ vs. other,
• the one true bracing style,
• javadoc style commenting vs. freestyle commenting vs. no commenting at all,
• SVN vs. git vs. other,
• design-first vs. happy hacking,
• the person calling the shots,

all at once?



My experience with Pro-POS: Do not take people on board unless you have a strong core of already-done work. Design by committee doesn't work, benign dictator works only if said dictator has the authority with everyone to be accepted as such - and a project set up from the beginning to be a community project cannot have such a person...


Edit: You're retarded...

*grabs the statistics*

#languages = 3 (C, C++, ASM)
#1TBS = 2 (Unix vs Allman)
#comment styles = 3
#version systems = 3 (svn/git/hg)
#coding style = 2
Total mindsets = 108
P(accept another as leader) = 0.25

People needed to likely have a common mindset = 11
People needed to likely start community project = 11 / 0.25 = 44

(Birthday paradox owns)

Edit: People needed to start a community project and keep it alive = a lot

In all honesty, nobody's ever gotten anywhere with a community project. Look at, for example, Comos. Lasted, oh, roughly four or five weeks IIRC. I didn't like the layout of the code or the choice of repo software OR the way the project was put together. The team "leader" wasn't that fun to work with (very pushy.) The project was a failure before it started, which was too damn fast. There was practically NO design put in, just bare bones, grub-booted, slowly-developed code.

Epic. Freakin'. Fail.

--Troy

DexOS started as a community project, it started on the v2_os forum, half the members want to go back to the old design and half wanted to stay with the new, but failed design, i had already been working on a ASM for about a year and a half.
Because we could not agree we, split and desided to code a new OS, which was based on the design of the old v2, but would be all new code, including change from tasm to fasm.
In the end not a single line of code was used from v2_os, but the good news is the origanal team is till togeather, that is over 5 years ago.
Each team member has worked on part of the OS, plus many new members.
The main point is theres many more things to a project than coding, theres site, wiki, doc, testing, support, tranclation, etc.
I am not say DexOS is sucsessfull, compard to linux and such lights, but as a team we have achived all the goals we set out when we started the project.
Once you have done that, moving on to the next stage is much harder.

But that is, effectively, a fork / rewrite, not a project from scratch. Different story...

Troy Martin wrote:In all honesty, nobody's ever gotten anywhere with a community project.

I beg to disagree. If you had said most projects, maybe. But "nobody ever" is a too strong claim.

Look at, for example, Comos. Lasted, oh, roughly four or five weeks IIRC.

Proof by example?

Kevin wrote:Proof by example?

Comos was a community-written OS project that was in the #communityos channel of freenode. We still use the channel, but the project failed after about a month.

Solar wrote:But that is, effectively, a fork / rewrite, not a project from scratch. Different story...

Thats like saying linux is a rewrite of MINIX, the only thing the v2_os as to do with it, is we met on there site, the OS had been started years earlier.
I just took my opportunity, to get a like minded team togeather.

I wonder why somebody hasn't come out with a more feature full version of minix....

nekros wrote:I wonder why somebody hasn't come out with a more feature full version of minix....

Because that would be an oxymoron.

But more importantly, I personally think a collaborative OS project could work... just not with THIS community .

The only way to successfully do that around here is to ring up the experienced OSDevers and get them interested enough to work on the project (difficult as it is, as you need a project that is inspiring and realistic enough, a way to ensure continued interest and proper collaboration, and to have already built a reputation for yourself in OS development). Also, try and avoid the newbies of the forum - they may be willing to learn (though this is becoming more and more untrue), but they will be more of a hindrance without previous OSDev experience behind them.

Zenith wrote:But more importantly, I personally think a collaborative OS project could work... just not with THIS community .

I think that is probably as correct as anyone is going to get. We all have different coding styles, and we all like different types of kernels and designs. Some of us are pro-POSIX, some of us hate it. Some of us like C, some of us like assembly. Some of us think community projects are a waste of time, some of us don't.

We here are too different to actively participate in the creation a large community-written operating system. Like I said before, semi-advanced person starts project, project attracts more semi-advanced people, project fails because members work on their own OS instead and actually get somewhere with it.

Guys,

A community OS can work, and can work with this community. The project I am part of, Pedigree, has four developers, all from this community (Myself, Pcmattman, OrOs, and AJ). One of the founding developers/designers of the project was also from this community (bluecode, although he has since left the project).

Our project seems to be coming along nicely. I think you're really thinking up problems where there are none. Bracing style etc doesn't matter, as long as there is a coding standard created that everyone adheres to. It's unlikely you'll match everyone's One True Style but most people can make compromises and still be happy looking at the code they create.

At least, if they want to survive in the business coding world they can.

Cheers,

James

nekros wrote:I wonder why somebody hasn't come out with a more feature full version of minix....

That was the intention of Linux. Heard of it?

As for the whole community projects business, I remember someone on this forum mentioning a pretty well-known project (can't remember which, though) that started out like that. If you have the patience, you can dig deeper.

Yeah, but linux isn't a microkernel and I'm talking about using the actual minix kernel and adding programs like gui and stuff like that. I kinda like the minix design, although admittedly I probablly wouldn't use it in daily life because of it's design purposes.