Re: Re: Debian BSD.. cool idea
On Sun, Jan 30, 2000 at 10:13:05PM -0600, Steve Price wrote:
> On Sun, 30 Jan 2000, Raul Miller wrote:
> # Anyways, given this supposedly wonderful support for linux binaries,
> It _is_ wonderful! Have you ever tried it?
It works fine on most applications, but it's missing a couple things
(syscall() support, last time I checked -- and I have a vauge recollection
of something else).
> # perhaps it could be built into something stable enough to support multiple
> # bsd kernel versions?
> I'm not sure what you mean here.
Please read the text at http://bugg.strangled.net/debbsd.txt
I was responding to the issues raised there.
> # Or is that as realistic as expecting FreeBSD to be portable across
> # multiple cpu architectures? [Did I mention I'm mystified that BSD folks
> FreeBSD (and *BSD) is portable across multiple architectures! I use
> FreeBSD on both x86 and AXP machines right now. There is a port for
> m68k and I'm working on an UltraSparc port as I type this. Before
> BSD became FreeBSD and NetBSD (and OpenBSD after the Theo blowup) it
> ran on a number of architectures. The four distributions of BSD
> doesn't even begin to add up to the 100+ different distributions of
> */Linux/*. The only reason FreeBSD doesn't have as many ports is
> because of demand and the goals of the project. It is more concerned
> about serving the majority of the planet with an OS that is the fastest
> and most stable one it can provide. That it runs on your toaster or
> microwave is something FreeBSD has left up to NetBSD (and Linux to a
> lesser extent).
Ok. I was hoping that the two forks (Free/Net) would unfork, but I've
been harrassing FreeBSD folk about the lack of portability off and on for
several years now. I guess I can live with a portable FreeBSD kernel,
but I'm still disappointed about the split.
> # criticize the design of the Linux kernel when NetBSD isn't as fast
> # as Linux and FreeBSD isn't as portable?]
> Please qualify your remarks since you seem intent on slinging FUD.
> NetBSD is intended to work on every architecture. NetBSD is concerned
> with portability first and speed later. FreeBSD is pre- dominantly
> tuned to run on the Intel architecture, though recent ports to other
> archs have proven to be just as stable and fast. It doesn't matter to
> me if FreeBSD runs on my grandfathers stopwatch. What matters to me is
> that it is fast, secure, and stable on the architecture that better
> than 3/4 of the world runs on - Intel.
> The comments you see about the design of the Linux kernel from the
> BSD community are undoubtedly not because of its speed or
> portability. Rather it is because the BSD kernel is ~30 years in
> the making and Linux has about 9. The BSD kernel was and contiues
> to be the basis for a number of commercial (including realtime) OSs.
> Linux is seeing more and more uses, but please don't discredit the
> decades of hard work that went into designing what is the BSD kernel.
> You are definitely entitled to your opinion, but please don't do it
> with baseless comments like this. Provide facts and real numbers
> and we'll all believe you. Otherwise you are just spitting in the
I'm unhappy about the Free/Net split. What kind of numbers do you want
from me to justify this unhappiness?
I think that, given that so many people think this split is justified,
that criticisms of the Linux approach are superficial. Why do you
characterize such comments as baseless?
It's good to hear that FreeBSD is moving in the direction of being
portable. Maybe this means that NetBSD can move in the direction of
I can hope.
[I've often wished that RedHat and Debian could merge too, by the way.
But rpm is still missing something of what we have in dpkg, and vice
versa. And there's nothing even approaching the sort of meta-conflict
database that we'd need for anything more than superficial overlap
between the two.]
Raul is responsible for 14 BSD boxes, at the moment.