[haiku-development] Re: What's the status of Haiku?
- From: Sia Lang <silverlanguage@xxxxxxxxx>
- To: haiku-development@xxxxxxxxxxxxx
- Date: Thu, 21 Aug 2014 19:14:15 +0200
On Thu, Aug 21, 2014 at 7:04 PM, Paweł Dziepak <pdziepak@xxxxxxxxxxx> wrote:
>
>> If it was at least some kind of modern, interesting kernel... then I
>> could be persuaded to join the train, but with an archaic (design-wise)
>> kernel like Linux... it really isn't convincing to me. I once had the
>> 'joy' of working myself deep into a few parts of the Linux kernel
>> sources, and what I saw was certainly not pretty. Working inside the
>> Haiku kernel was a great experience in comparison!
>>
>
> That's something hard to agree with. Linux is the place where most of the
> innovations in the world of kernels take place, e.g. various versions of
> RCU, fair scheduling, network channels, zswap/zram, etc. What is more
> modern than that? Obviously, writing object oriented code in C results in
> the necessity of ugly tricks (like container_of), but that's more like C vs
> C++ discussion and there is nothing archaic in that.
>
> Paweł
>
Yes, exactly. With the proposed setup, the vast majority of Haiku work
would be in BeOS-API-land, not Linux. The "nice" C++ code lives on. Nothing
changes except a great kernel with the driver situation solved.
(The added complexity for adding drivers is there for good reasons (power
mgmt for one), but then again, other people are already writing these
drivers.)
Sia.
Other related posts:
