And that's what I'm saying is the correct approach. You first rewrite some basic things for Linux to run as a server. The micro kernel needs to be written from scratch, obviously, since it's a completely different design. The thing is, when you write a micro kernel, you don't just write the micro kernel, you write a set of servers providing functionality. If you start ALL THE SET from scratch, you will get a virtually non-advancing project, such as Hurd is. If you start with the kernel, and take a full fledged one as a server, you can do the rewrite gradually all of the features of the Linux kernel as smaller servers, without loss of functionality in between. This way, you get real usage testing, since it still offers support and features.The Fork Linux aproach has some problems, first I dont think its easy to rewrite a monolitic kernel to a micro-kernel, often in such big refactorings its easier to rewrite stuff, but yes maybe it would be better.
Too unfree? The problem with GPLv3 is as simple as the fact it's almost impossible to not step over someone's license, and nobody wants to take the risk of being responsible of it. That's the only "advantage" when using GPLv3, that the developer is the one to sue if it steps on someone's patent.2nd I dont want to develop such stuff so it does not really matter how you come to a good micro-kernel, another problem is that the linux kernel is to unfree, so if you start from a freebsd kernel you get patent law suits like hell, with linux u cant fork your code to normal gpl, just to a modified version (v2-only).
As for the FreeBSD, it's the same story as with Linux, and it's kind of the same with GPLv3 software, with the only difference that the one paying will be the developer. Really fair, huh?
Even Stallman admitted Linux does good as the GNU kernel. Oh, well...So to go the gnu way would be the best, to get 2 things done, its not very effishent to rewrite a complete kernel to only fix 1 of 2 problems. If you do that work fix both.
Also, writing from scratch a kernel AND a lot of servers is a lot more work and a far less efficient than just rewriting the most basic part of a kernel to a micro kernel and refactoring a bit of the Linux kernel to allow it to run as a server. You seem to disregard testing and support. That mentality is the one which leads Hurd to be stagnated.
And you can gradually drop all of it. You can upload any changes on the servers and your micro kernel with the license you want. GPLv2 is good enough for having something working until you have a full rewrite of the functionality.I mean if you would say 3 and later, somebody could fork it if gpl 4 would suck and comply only to version 3 rules so I dont see any disadvantages in that, on the opposite end you cant upload changes in gplv3 if you like your code protected by that. Even GPL2 is bypassed by DEFINE_KERNEL_BLABLA_DO_WHAT_YOU_WANT_ITS_NOT_GPL symbols.
And again with Android. For a lot of users, speed matters. Also, using a java API leads to better reliability than native code (because most programmers suck at managing memory, mostly because they are lazy and they are expected to work fast, not to work right), because it runs on a VM. With your logic (speed doesn't matter), Android can not be a piece of shit, but rather brilliant. I don't know what "akkus" means.Android is a piece of shit, and if we talk about speed android is the best example that for most people (most people arent linux network admins, but most have a smartphone) speed does not matter the java-api makes android so slow and especialy ram-hungry that you need basicly double the memory and double the cpu speed to have the same speed than a apple or microsoft phone needs. That leeds to the need of bigger akkus...
Most users think cellphones are slow, and just that. They don't expect real speed from them. They do expect to be able to get some performance of a desktop. I don't think Hurd is targeted to cellphones.BTW I never said that NOBODY cares about speed, of course some geeks do, but most people dont (android). btw I dont know about apples smartphone os, but its a hybrit? kernel or something and liek a said 1000x faster than garbage java + linux in android.
Also, the slowness of Android is due to the user space, not due to the kernel. I don't know which kernel iOS uses, but OS X uses an hybrid, with FreeBSD as a server on top of a micro kernel. The main reason it's faster is because they use native code, and there is no garbage collection. Again, user space reasons.
Who said licenses doesn't matter, again? As far as I saw, the only problem with Linux would be that the developer does not need to be sure his code is not under any patent. Assuming he would sue you because he owns a patent covering the code he freed up would not be smart, IMO, because the solution would be to just revert a patch and he'd get nothing, so the only consequence of GPLv3 would be that if it's covered by patents, he will get sued. And chances are for any given algorithm, that either you own a patent for it, or others do, and you'll get sued instead of the ones benefiting from your work. GPLv2 is still GPL. Also, if you have a problem with licenses that hard, I expect you to not use a GUI, because both Wayland and X.org use the MIT license, while Mir uses a CLA that gives Canonical the same rights the MIT license gives anyone else.I could even live with a gpl2 + later kernel... If I would make changes they would be released in gpl3 and if the new linus would not pull that, ok I would have no big problems with that, if you comply to gplv3 u get all stuff from gplv2 the companies that dont want to comply to gplv3 just get the hole deal their problem.
AND one nother thing if lisenses would not matter everybody would switch to bsd because its cleaner and better than linux.
And BSD is not cleaner and better than Linux, AFAIK. BSD is lacking a lot of support, you know?
Can't you see there is no practical difference between writing from scratch and rewriting, aside from being able to use already implemented functionality with the rewrite?And that will not change for the next 5-10 years maybe even 20 years. But maybe Hurd is really dead (even Hurd was never = Hurd it was different oses as far as I read it, so which hurd in the future is now dead would be the question) but I just say, linux is about rewriting everything why should the kernel be the only thing where this will never happen?
Linux is GPL FOSS.Replace Linux with GPL FOSS Operation systems. or "GNU"