Newsgroups: comp.unix.bsd
Path: pa.dec.com!decwrl!uunet!uunet!karels
From: karels@BSDI.COM (Mike Karels)
Subject: Re: Jolitz: The Road Not Taken
Message-ID: <1992Mar23.214834.7381@uunet.uu.net>
Summary: Roads that were taken
Keywords: 386BSD, Jolitz, motivation
Sender: usenet@uunet.uu.net (UseNet News)
Nntp-Posting-Host: bsdi.com
Organization: Berkeley Software Design, Inc.
References: <579@svcs1.UUCP>
Date: Mon, 23 Mar 1992 21:48:34 GMT

I wish it were not necessary for me to reply to the recent posting made
on behalf of Bill Jolitz ("The Road Not Taken"), but given the nature
of the comments about me, my past relationship with Bill, and my work,
I am compelled to respond.

I do not understand Bill's complaint with CSRG, with me or with BSDI,
nor do I understand why he is now saying the things that he is.  The
recent posting is not at all an accurate reflection of past events.
I do not want to correct it point by point, as I will not continue this
argument in a public forum.  However, there is nothing for which I should
apologize.  I have not lied to Bill or other friends, and I have not
asked anyone to act unethically or immorally.

Bill was a founder of BSDI as well as its first full-time paid employee;
he can hardly be surprised that BSDI is selling a supported system
based on the freely available Berkeley code, of which he contributed
the 386 port.  I did not exert pressure on Bill to work for BSDI,
although I did introduce him to the other founders.  When he became
unhappy working for BSDI, I had a number of long conversations with him
about the problems.  Although I never fully understood the problems,
his complaints centered on business and personal relationships within
the company rather than the fact that the company planned to charge
for its product.

Bill's complaints about CSRG are unjustified, and only originated in January
after I decided to leave Berkeley to work for BSDI.  CSRG released all
of the code contributed by Bill in source form.  His complaint is the
lack of a supported binary release from the University.  However, Berkeley
releases have never been supported in the normal commercial sense, and
have never been packaged and documented for easy installation on machines
as diverse as 386 AT systems.  While I was in CSRG, we never considered
doing another binary release after NET2 until the alpha release of 4.4BSD.

Although the NET2 release contained most of the BSD kernel, several
critical modules were missing because they were derived from licensed code.
CSRG could have taken the shortest path to filling in those modules,
but those modules were among the oldest in the kernel.  Rather than
reimplementing those pieces as they had been, CSRG chose to redesign
them properly, which is in progress.

I regret that this disagreement was made public at USENIX and in
this news group.  I will not continue this discussion in public.

		Mike Karels
		Berkeley Software Design, Inc.
		karels@BSDI.COM