On Thu, 11 Aug 2016 11:03:13 +0200
Anselm R Garbe <garbeam_AT_gmail.com> wrote:
Hey Anselm,
> I have no problem with OpenBSD per se, but I do think that its scope
> is general (server) purpose and that the hardware support in the
> embedded non-network space doesn't sound too great to me. linux kernel
> + some dedicated userland is far more flexible imho.
hardware support is not as good as we see with Linux, but consider how
much money flows into Linux compared to OpenBSD and how many companies
push changes into the Linux kernel.
And things are improving and there are numerous laptops and desktops
running very well with it. Given how dedicated the suckless fellows are
anyway (the only possible userbase for stali tbh in the short term),
I'm sure they'd even invest into a laptop or desktop with the special
interest of it supporting OpenBSD.
Another thing is: Look at the OpenBSD drivers, they are designed like
space-ships. Compared to this, Linux kernel code often looks like a
racing kart held together with duct tape. I'd rather use a system
offering less hardware support but genuinely better security and
stability than some clamped on solution.
My thoughts are this: We could either go ahead and create yet another
Linux distribution with, even if it was successful, yet another
fragmented community. Package maintenance is not an easy task, and as
we discussed at the last slcon, you intend to rewrite all the build
systems for the different projects to make them suckless, which is
ambitious but also too ambitious for a project with such little
manpower.
We need to save our energy for projects that matter. The OpenBSD ports
system is solid and well-maintained; we have strong arguments on our
side to justifiy why static linking is in many ways more secure, stable
and lightweight than dynamic linking.
OpenBSD has some old cruft of course, but with Linux we are just
scratching the surface. The entire userland is a toxic environment.
The OpenBSD developers are sane guys and have complete control over the
userspace; they discuss with reason and are not frightened to cut big
chunks out of the OS if they see fit.
The static approach of course would not apply to all packages, but I
wouldn't even be surprised if they accepted such changes for certain
things.
As an example: Just look how easy it is to setup Wifi on OpenBSD, or a
RAID, and many other things. This is years ahead of Linux.
I am sure suckless.org has a lot of street-cred in the OSS-scene. We
could use this leverage to have a positive influence on a big
distribution people actually use. In the long term, making OpenBSD
better will benefit those who are scared of making the switch. We all
know that OpenBSD is much further on the convergence line towards an
ideal operating system for server and desktop applications than Linux
and its messy userland.
Cheers
FRIGN
--
FRIGN <dev_AT_frign.de>
Received on Thu Aug 11 2016 - 11:21:59 CEST