*BSD News Article 2715


Return to BSD News archive

Path: sserve!manuel!munnari.oz.au!spool.mu.edu!sgiblab!zaphod.mps.ohio-state.edu!magnus.acs.ohio-state.edu!usenet.ins.cwru.edu!agate!mastodon.cs.berkeley.edu!eric
From: eric@CS.Berkeley.EDU (Eric Allman)
Newsgroups: comp.unix.bsd
Subject: Re: I wonder, did AT&T backstab BSDI?
Keywords: AT&T speculation lawsuit BSDI
Message-ID: <157693INNrkk@agate.berkeley.edu>
Date: 29 Jul 92 22:32:03 GMT
Article-I.D.: agate.157693INNrkk
References: <1992Jul28.153750.8395@mintaka.lcs.mit.edu> <1992Jul29.144859.8222@bas-a.bcc.ac.uk> <1992Jul29.174437.18606@gateway.novell.com> <1992Jul29.201919.15968@kithrup.COM>
Sender: eric@mastodon.cs.berkeley.edu (Eric Allman)
Organization: UC Berkeley Mammoth Project
Lines: 77
NNTP-Posting-Host: mastodon.cs.berkeley.edu

In article <1992Jul29.201919.15968@kithrup.COM>, sef@kithrup.COM (Sean Eric Fagan) writes:
|> In article <1992Jul29.174437.18606@gateway.novell.com> terry@thisbe.npd.Novell.COM (Terry Lambert) writes:
|> >Things AT&T owes to Berkeley:
|> >2)	VFS
|> >3)	Memory management strategies
|> 
|> That's funny, I could have sworn that these actually came from SunOS.
|> Considering that 4.4ish is the first BSD system to have vnodes, and a decent
|> memory-management scheme, I don't see how SysVr4 (which has been out for a
|> couple of years now) could have gotten it from 4.4...

Vnodes did indeed come from SunOS, although I understand that Kirk McKusick
and Mike Karels contributed to the design.  However, memory management was
developed jointly between Sun and Berkeley -- Kirk McKusick spent
considerable time on the design.

|> >4)	Job control
|> 
|> Other systems had job control long before BSD did.

Jim Kulp first put job control into a PDP-11 system at IIASA in the distant
past.  It was derived from (I believe) ITS, although my memory is a bit fuzzy
on this.  He donated it to Berkeley, where Bill Joy merged it into BSD.  It
may have come out first on a non-UNIX system, but AT&T/USL clearly got it
implementation-and-all from Berkeley.

|> >5)	csh
|> 
|> I use bash or ksh.  csh has too many problems.

Perhaps -- but it is still a common shell, and it is certainly another
example of a program developed at Berkeley picked up by AT&T/USL.  And
csh certainly introduced ideas that still exist (in possibly different
forms) today.

|> >8)	mail/sendmail/smtp
|> 
|> Other systems had smtp before, no?  sendmail isn't that great, and there are
|> other systems that deliver mail without the horrible configuration problem
|> that sendmail is.  And UNIX had mail *long* before UCB ever started playing
|> with unix!

There are a lot of opinions about sendmail, and I agree with many of them.
Yes, there were other programs that delivered mail without "that horrible
configuration".  /bin/mail delivered UUCP mail.  Another local mail program
done at Berkeley delivered "BerkNet" mail (an RS-232 network, now gone,
written by Eric Schmidt).  And the NCP FTP code had yet another mail program
that knew about NCP mail.  I wrote delivermail so that users didn't have
to decide which mail program to use depending on what network their mail
was going to.  It had a very small compiled in configuration file.  It was
renamed "sendmail" during a major rewrite that pulled the configuration file
out; the changes in sendmail were adaptations needed to handle the real world
that was changing fast (much faster than now).

SMTP didn't even exist before sendmail.  The conversion from delivermail to
sendmail occurred during the network conversion from NCP to TCP.  SMTP was
in draft state (I probably still have many of the drafts in my basement).
When a new draft came out I would FTP it and usually have the changes
implemented and installed by the following morning.  RFC822 was similarly
evolving.

Yes, "UNIX had mail *long* before UCB ever started playing with unix!"  It
was essentially the "cat" program.  It didn't know about networks (later
versions did know about UUCP, but that appeared well after UCB started
playing with UNIX).  It didn't even understand about separate messages;
reading your mail was literally equivalent to catting the mail file.
Kurt Shoens wrote a mail UA while at Berkeley (a derivative of which is
usually known as mailx) that understood about separate messages, replying
to messages, headers, and all sorts of "new" ideas.

|> -- 
|> Sean Eric Fagan  | "My psychiatrist says I have a messiah
|> sef@kithrup.COM  |  complex.  But I forgive him."
|> -----------------+              -- Jim Carrey
|> Any opinions expressed are my own, and generally unpopular with others.

eric allman