Technology and Trends
 USENET Archives
  
Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!utgpu!water!watnot!watmath!clyde!cuae2!gatech!lll-lcc!lll-crg!hoptoad!gnu
From: g...@hoptoad.UUCP
Newsgroups: comp.mail.uucp,comp.sources.d
Subject: Ownership of uuslave.c -- particularly, does AT&T have any ownership?
Message-ID: <1800@hoptoad.uucp>
Date: Fri, 13-Feb-87 16:40:38 EST
Article-I.D.: hoptoad.1800
Posted: Fri Feb 13 16:40:38 1987
Date-Received: Sat, 14-Feb-87 23:08:34 EST
Followup-To: comp.mail.uucp
Organization: Nebula Consultants in San Francisco
Lines: 71

[Followup-To ==> comp.mail.uucp]

I have heard a rumor (first-hand and now second-hand) that AT&T is
gearing up to "crack down" on license violators re Unix sources.  
(I think this is probably a good idea for them, else they'll lose
all rights in what they have.)  However, the rumor continues that
in particular, they are tracking what people say & do with uuslave.c.

So far if this is true, it is only a rumor -- nobody from AT&T has come
out to the net and said "I believe uuslave.c is covered by your Unix
license and is therefore not public domain".  In the absence of such a
claim, and with evidence that it is *not* AT&T derived code, I am forced
to assume that it really is public domain.

The BBS system that it came from (which is so hot and heavy on verifying
who the people are who log into it to DOWNLOAD) cannot tell us who UPLOADED
uuslave.c -- they threw away their records and didn't keep a backup.

I have checked with several very technically competent people and asked
them to read uuslave.c and compare it to the Unix uucp sources.  They
all say "I see no correlation, this is definitely not ripped off
code".  I encourage anyone else out there in "netland" to do the same
and post your results to comp.mail.uucp.  There is a single subroutine
for computing the checksum of a packet, which may originally come from
uucp, but it was published in Greg Chesson's "Packet Driver Protocol"
paper, which was given out to anyone who asked for it while Greg was at
Bell.  (I recently posted this paper to comp.mail.uucp.)

In reading the code, and working with it, it's clear that it was not
written by someone who really understood packet switching protocols and
acknowledgements.  The code didn't even check the packet checksums,
though of course it had to generate them on outgoing packets!  It was
probably done by someone who had access to the protocol description
(which doesn't say *how* to implement the error checking and such, just
says what all the bits are in all the packets) and coded it up and
tried it against a Unix system.  Whatever packets came back at them,
they wrote code to expect at each point.  This doesn't make for a
robust implementation, so I have rewritten this stuff anyway.  But it
shows that the author(s) were probably not working from uucp sources,
or they would have structured their code more like the uucp sources.

Note that Lauren Weinstein has successfully reimplemented uucp without
looking at the Unix sources.  His product is now sold commercially for
MSDOS machines.  This proves that it can be done; AT&T can't claim that
"nobody could do it without looking at licensed source, therefore we must
have *some* right in it".

I'm not going to stop working with the code based on rumors and scare
tactics.  I just wanna get the facts, ma'an.

I hereby call upon AT&T, the company and its employees, to come out and
state whether public use of uuslave.c is believed to violate any of
AT&T's rights (eg copyrights, trade secrets) or licenses (eg Unix
source licenses).  Please state the reason for your claim, if you do
claim a violation.  In the absence of such a claim by AT&T, I will
continue to assume that the code is public domain, and that even in the
unlikely event that there is some AT&T involvement in uuslave, the
company has decided to abandon any claim it has in the program.

AT&T employees, please forward this to the licensing department or whoever
handles these kind of questions.

	John Gilmore
	Nebula Consultants
	PO Box 608
	San Francisco, California, USA  94115-0608
	+1 415 931 4667

-- 
John Gilmore  {sun,ptsfa,lll-crg,ihnp4}!hoptoad!gnu    g...@ingres.berkeley.edu
			"Use the Source, Luke...."