gabriel rosenkoetter on Wed, 29 Jan 2003 00:26:09 -0500


[Date Prev] [Date Next] [Thread Prev] [Thread Next] [Date Index] [Thread Index]

Re: [PLUG] directory conventions


Oh, fine, I'll bite. :^>

On Tue, Jan 28, 2003 at 07:24:46PM -0500, William H. Magill wrote:
> Solaris is more System V than BSD with no OSF added. The old Sun OS was 
> similar to BSD until Bill got cosy with IBM and the resulting Solaris 
> looked mostly System V-ish.

Uh, Solaris is very purposely System V and the old Sun OS (anything
pre 5.6) was very purposely BSD. As in, Sun said that's what they
were trying to do.

(Never you mind /usr/ucb, silly user. You're not suppoed to know
that exists anyway. If we wanted you to know what it was there for,
we'd have called in /usr/bsd, and if we wanted you to use it, we'd
have made /usr/ucb/ps execute in something resembling a sane period
of time.)

> Digital Unix, aka Tru64 Unix was the only version of Unix to ever 
> actually implement the OSF code. Other vendors added features to make 
> their proprietary versions "OSF Compliant"

Whoop. OSF. Cry me a river. ;^>

> And then BSD itself isn't very BSD like any more. And Darwin is?

More than OpenBSD, FreeBSD, and NetBSD, arguably. It's 4.2 BSD
based because it's NeXTStep-based... and NeXTStep is painfully BSD
everywhere it isn't painfully NetInfo. (Go on, ask me why I don't
boot my NeXT turbocube much. Okay, I'll tell you: a friend was using
it on Swarthmore's campus for a while, and since the disk is small,
configured a lot of things, including a lot of non-essential, though
oft-used, libraries on the other end of an NFS mount. But I can't,
for the life of me, get far enough into NetInfo without the whole
thing hanging on an NFS read that will simply never work to switch
those mounts OFF.)

Then, of course, they hired FreeBSD people to spruce it up.
Whatever.

> As for "direct member," it depends upon which branch and how much 
> incest you want to allow for  - USG vs OSF and all that good stuff. Now 
> AIX is more System V than BSD with a measure of OSF thrown in.

I just mean "source is based on something that had some real right
to be called Unix at some point". And by "based on" I don't mean
"behaves like", but "is the product of a function performed on the
original code".

> Linux is most definitely NOT a Unix(tm)... because Unix(tm) is a brand 
> name and there is an entire suite of tests which must be passed before 
> one gets to pay the fee that allows one to call your thing -- Unix(tm). 
> And you'll never catch Linus trying to pretend it is a Unix or calling 
> it one.

Linus has a very good concept of just what his OS is and isn't
(well, um, duh :^>). But there are plenty of people out there who
don't.

> Some go so far as to claim -- if it doesn't respond to "man hier" it's 
> not "really" Unix. (hier is the description of the "Standard file 
> system hierarchy").

I'd agree with that in a second. I don't think it's going
particularly far. (Now, if it doesn't have hier(7) because MANPATH
is broken or you didn't install the manual, that doesn't count. ;^>)

I don't know that having hier(7) means you're Unix-derived, though.

> /local is a new invention. Before that it was /usr/local and before 
> that it didn't exist

Not sure how you mean you see /local being used... at
cs.swarthmore.edu, /usr/local is locally built software that's NFS
shared from the server to the workstations in the lab and /local is
the overflow of the (way too big for just the OS) truly local hard
disk. I'd had the impression we were following some kind of common
practice in doing that (and I maintain that it's a sane
interpretation, even if that's not the case ;^>).

> And then of course is /usr/opt... :) [Dammed if I know, some vendors 
> use it some don't.]

Or even just /opt, which I don't see in *my* hier(7), but I'm
reading a BSD one. ;^>

I blame commercial vendors (who are also responsible for today's
gross abuses of LD_LIBRARY_PATH).

> Next we can go argue Unixness on Computer Science grounds -- is a 
> micro-kernel or a monolithic-kernel "Unix."

Oh, either or. It depends on whether you're running it to think
about OS design or running it to get real work done. ;^>

> By the way, I counted 723 on the head of my pin...

Heh. That's funny, I always find mine are in powers of two...

-- 
gabriel rosenkoetter
gr@eclipsed.net

Attachment: pgp0qPKmx8YUG.pgp
Description: PGP signature