PureDarwin - Community project to make Darwin more usable | Hacker News Reader

PureDarwin - Community project to make Darwin more usable

puredarwin.org

47 pointsics30 comments

No blog updates since Dec 2012, no forum posts since April...

https://groups.google.com/forum/#!forum/puredarwin

Obviously nobody is interessted. Not because the system is bad, because Apple user want to use Apple and not an Open Source alternative.

It make more sense to create a good emulator, so you can use OpenSource System (like Debian, BSD,...) and native Mac OS Software.

I would be thrilled to have a solid Darwin that runs on commodity hardware so the servers could run the same code as the clients. The last I looked at actually running Darwin it looked like there was no effort from Apple to maintain a runnable Darwin. The lowest risk option was writing the code twice, a waste of resources we could have done without.

Which servers and which software? I'm curious. Most of the stuff in that realm is platform independent, anyway.

I want to run my own daemons on the server. But I want them to be written in Objective-C and use the Foundation libraries. This is all part of Apple's Darwin released code, it just doesn't run on anything.

I want to have access to the instant scalability of the various VPS and dedicated box providers, but with Apple ecosystem code.

Apple abandoned the datacenter hardware market long ago, unfortunately, they took the datacenter deployment options with them.

How much of the system is this even? A kernel is not really enough to run software on, not easily at least.

I don't understand the motivation for this project. OpenDarwin (yes, not PureDarwin) was motivated by putting Darwin on Intel, and when Apple moved to Intel processors, it pretty much killed OpenDarwin's steam. I get that people have their pet projects and do things "because they can," but attempting to drum up support for your project from the community requires that you explain to people why they should care.

Some people are interested in learning how things work. In that case the main page's description is enough to get them going.

I don't think that every project needs to be treated like a commercial product. This kind of project usually doesn't have a direct value to consumers but prepares the ground for other projects. It can help better understand the OS and help jailbreak IOS. Or provide drivers to install OSX on a PC or a VM and then allow cross-OS browser testing. Or maybe help a wine-like project for OSX apps. Who knows but without that project all these might not happen because the barrier to entry might be too high.

And this is better than Debian, Ubuntu, Mint etc. because -- what?

It's usefulness is source level compatibility with Mac OS. It's a tool for understanding Mac OS, testing Mac components in a detached fashion, redistributing Mac software without proprietary baggage or running a pseudo Mac environment on commodity server hardware.

Even as a standalone OS, a comparison with other BSDs is far more appropriate.

Hell if you could somehow pry out OS X components compatible with the build, it might run a basic OS X clone. There'd probably be a lot of hwid spoofing involved though.

Yes, I was wondering about using it for testing some (very low level) osx stuff for automated testing in a vm. But thats about the only use I can think of...

That proprietary baggage is what makes Mac OS X worthwhile as an OS.

Without Objective-C and all the frameworks that make up the OS, it is no better than using a *BSD.

A lot of people would disagree with you, obviously. That's a purely subjective statement disguised as objective analysis. I'm not saying I disagree with you, but it's a complex enough issue to merit a heck of a lot more thought and research than you have seemingly put into it.

> Without Objective-C

GNUstep.

> it is no better than using a *BSD.

Actually it's quite likely to be worse. Is there any good reason Apple is using that Mach stuff other than momentum and having developers who know it well? IMO any given BSD project has a more coherent design than XNU.

GNUStep is almost dead project, it was being developed already before Apple bought NeXT, and it is still not there.

The way the web site gets updated, looks like a software project connected to a machine, without proper support for the latest Objective-C versions, who would bet on it for production code?

Every time I see them at FOSDEM I look forward to see better updates, but it always looks the same.

I used to use and develop with GNUstep (just small personal projects) for a good chunk of the 2000s decade, until I lost interest. A lot of the important stuff is there. AFAIK they have parity with 10.4 or so.

It's true that there isn't a lot of momentum behind the project. I suspect that's just a lack of people. I wonder if they'd be receptive if some folks just jumped in and implemented more of the missing stuff. I am guessing yes.

The spirit of my comment is: you can't say objc doesn't exist outside of Apple. It's there in some form. It might even more "there" if people start using and developing it. We as individuals have the power to do that. Dismissing it very early strikes me as defeatist.

> you can't say objc doesn't exist outside of Apple

Sure, it exists in a form as if we were in 2005, away from all Objective-C 2.x goodies, more recent frameworks and tooling.

... which is not nothing. These developments come from somewhere, I don't understand why one would bash GNUstep because it's not a unicorn.

I am not bashing it, just stating the hard reality that the project is stagnated.

BSD-licensed, not GPL? For some people that's important.

It doesn't quite have to be, from a user perspective. This is the kind of hacking project where just making it work well (boot + drivers + services) is a fun challenge. Further, it's a sandbox where you can play with some of the neat ideas that are at the heart of Darwin, like XNU.

Would it be feasible to have Debian + Darwin? I am wondering about driver availability.. Does Darwin support all hardware FreeBSD does? Is it stripped down to Mac-like hardware?

And in general what would i get from using Darwin as opposed to Linux or *BSD apart from the ability to run a limited amount of Mac-only commandline tools?

Also, looks like a dead project..

Also, given that Apple moves away from open source more and more [1], i really wouldn't go down that road... you'll end up surrounded by proprietary software, extensions, drivers more likely then not.

[1] https://plus.google.com/103674611711666213369/posts/7j1c2bZ7...

> Does Darwin support all hardware FreeBSD does? Is it stripped down to Mac-like hardware?

I think you have the common misconception that Darwin uses the FreeBSD kernel. It does have components from FreeBSD, and one of the principal engineers was the founder of FreeBSD. However, the kernel is actually Mach+XNU (Mach is a micro-kernel, XNU is a Unix-kernel-like wrapper around it).

This is just being pedantic, but XNU is the name of the whole kernel, which describes the combination of Mach and the BSD interfaces.

The reason interest is going to be so low is that, generally speaking, it is pretty easy to just have software that runs on both Linux and OSX, cross compile for both, an use the Linux kernel instead of trying to use a binary compatible kernel... which lacks anything resembling a proper environment.

Reminds me a lot or ReactOS http://en.wikipedia.org/wiki/ReactOS

AFAIK, this project is dead for years

The long-dead project is "OpenDarwin".

As the index page says: "(some people think of it as the informal successor to OpenDarwin)"

This project has been "dead" for about 6 months or so. Personally I don't thing that should matter much– it's an interesting project, and the more exposure it gets the better chance it has at being rejuvenated.

Reply on news.ycombinator.com