🚀 go-pugleaf

sclawson@sal.cs.utah.edu (Steve Clawson) wrote:
]      I can't believe how much of this crap is getting posted here.
] Sheesh.  Linux was out at least 6 months before I ever heard anything
] about 386BSD (at least the initial Jolitz free release).  I know this
] because I was running Linux at the time that 386BSD came out...anyone
] remember Linux releases 0.001 and 0.10 (when you still had to have
] minix to create a filesystem)?  Hell, anyone remember the minix-386 by
] Bruce Evans?

Yeah, he's that BSD guy, isn't he?  8-).

]      Linux became popular because it provided people with (not
] necessarliy in any sort of order, or complete):
] 
] - virtual terminals.
] - shared libraries (disks weren't nearly as big back then).
] - VGA console support (for things other than 80x25).
] - ran in less memory (1-2MB).
] - didn't crash as much.
] - had regular releases (no weird patchkits to install).
] - merged vm/buffer cache.
] - wasn't 'contaminated' (remember the USL - BSDI suit?).
] 
]     All of these things gave linux a tremendous head start.  However,
] both the NetBSD and FreeBSD groups have been doing a amazing job of
] getting their systems caught up and even surpassing linux in these
] areas and others (does linux have a ccd device driver?).  And they're
] stable too. =)

386BSD had virtual termials almost immediately after the 386BSD 0.1
release.  The 0.0 release occured before Linux could boot multiuser
and was still posting in the Minix group without a group of it's
own (in fairness, BSD was using comp.unix.bsd, which didn't
appreciate the Intel BSD people any more than the Minix group
appreciated the Linux people).

Joerg (not Joerg Wunsch?) had SVR3 shared libraries -- the kind
Linux originally had -- in BSD in October of 1993.  Months before
Linux.  They weren't integrated because SVR3 shared libraries
are inferior; didn't make them unavailable, however.

The "VGA console support" for non-80x25 display is fluff, but is
easy to do.  Linux did have it first.

I booted an ALR system with the networking optioned out in 640k.
WSU was running 386's with 2M as X termials.  So much for less memory.

"Didn't crash as much" is subjective.  The original Linux didn't
act any more stable than 0.1 on the AT&T WGS boxes I ran it on
(the things CMOS lies about the size of their memory).

The 0.0 and the 0.1 releases were reasonably spaced.  Subsequent
releases didn't happen because Bill (Jolitz) had family problems.
If several of Linus' relatives had died too, it would have delayed
several Linux releases.

As to "the wierd patchkits", I resent that.  Since I *wrote* the
original patchkit (and the original FAQ).  There was nothing wierd
about them.  They were add-on's to the 386BSD 0.1 release and were
*intended* as add-ons, *not* releases.  And since they supported
logical dependencies and order serialization, it beats the hell
out of the context diffs that were flying around Linux at the time.

Actually, the original patchkit scripts and tools are currently in
use by at least two commercial software vendors (that I'm aware of).

The Linux VM/Buffer cache is (still) only partially merged.  Try
doing file I/O on an mmap'ed file and testing data interaction
(I have; at least I run the software before making calims).  A
"merged VM/buffer cache" doesn't only mean "competing in the same
memory pool".

Finally, BSD was never "contaminated".  USL claimed it was, but
never proved their claim, and in fact settled out of court when it
looked like the decision was about to go against them.  You are
simply attempting to sow uncertainty (and it's not working).


Those who revise history are doomed to repeat it.


                                        Terry Lambert
                                        terry@cs.weber.edu
---
Any opinions in this posting are my own and not those of my present
or previous employers.