kurt@fluke.UUCP (Kurt Guntheroth) (04/04/84)
How can AT&T put a clause in their contract making it illegal to disseminate
any knowledge gained from looking in the sources? You may be pretty scared
of legal action until you consider the following item: UNIX[1], with sources,
is available to universities all over the country. Students frequently
study UNIX sources in classes. Students never sign contracts agreeing not
to use any knowledge gained in a product. It would mean nobody could write
an operating system who had taken an operating systems class. Silly.
Now, how could AT&T know if you had learned about sources at a univwersity
as opposed to learning about them at work? Still, I wouldn't go copying
sources wildly. At least I don't live in mortal fear of disclosing
something[2].
-----------------
[1] UNIX is a trademark of something that used to be Bell Labs, (AT&T?)
[2] I don't have access to UNIX sources anyway
--
Kurt Guntheroth
John Fluke Mfg. Co., Inc.
{uw-beaver,decvax!microsof,ucbvax!lbl-csam,allegra,ssc-vax}!fluke!kurtron@brl-vgr.ARPA (Ron Natalie <ron>) (04/06/84)
Wrong. It is illegal to discuss UNIX source code in class. If someone is doing this they are violating their eduacational license agreement (at least the agreements for V6, PWB, and V7. Thats how long it's been since I was bound by an educational licence). When we worked with UNIX we were required to sign a nondisclosure agreement before we were allowed access to the kernel sources. There was a rather big furor a few years back when some university started teaching a course in operating systems with Lions book as text. -Ron
kevin@unisoft.UUCP (04/08/84)
One: It has been my understanding that the Lyons (sp?) book published
in Australia was generally available in Australia. In other words,
I could walk into any bookstore and without flashing my AT&T V6
software agreement, buy a copy of said book. Now it seems to me
that if this is so, the trade secret on the V6 kernel would be
pretty well shot. I doubt that they confiscate those books when
you enter the US through customs.
Two: However, it is also my understanding that the book only contained
the kernel sources and not the utilities. The kernel is of somewhat
limited use without the utilities.
Three: It seems that AT&T's distribution of the UNIX sources
to universities was a two-edged sword. It surely got them market
penetration at a time when they were under the consent decree. On
the other hand, I'd hate to go into court as AT&T and have to explain
why I thought the OS deserved trade secret protection after giving it
to an institution dedicated to the dissemination of ideas. In addition,
my experience with universities has been the same as I've heard documented
here; that being that the professors teach it without recourse to a
nondisclosure agreement and everyone leaves with a bootleg source tape.
Four: I believe that the AT&T software agreement states that the
signer is under no duty to keep the information proprietary should he/she
obtain the information from a third party without the breach of the
agreement. This is pretty standard trade secret language and would
probably be implied even if it weren't in the agreement.
Fifth: Please correct me if I'm wrong, but I believe that there are
people on this network who have binary-only licenses for UNIX. That
is, they have never received source from AT&T. Now, the System V
customer agreement for binary licenses contains no nondisclosure
provision (a change from System III). The sole protection AT&T asks
for is a contractual promise not to copy the binary except as necessary
for use on a single CPU. So:
Fifth.1: What happens when one of the binary only types taps into
this network and gets an eyefull of some transmitted source?
Fifth.2: Since most of the new UNIX systems on non-DEC processors are
not optimized, it would be a fairly straightforward thing
to disassemble it and pair the disassembled code up to
the namelist. Most of the C language structures seem to match
up to some pretty distinctive assembly code. This approach
would get you the utilities fairly quickly. Finally, you
could get a fair ways along the road of documenting the
sources by reading the Lyons book.
Does anyone have any better guesses on these issues? As usual,
UNIX is a trademark of Bell Labs (but who cares?)
Kevinfcz@pur-phy.UUCP (Fritz C. Zuhl) (04/25/84)
Paraphrasing >Students never have to sign anything that prevents them from disclosing >the source code of UNIX. I am student that must use UNIX for course work, and the whole class ( about 200 students ) had to sign an agreement that we would not disclose the workings of UNIX. I think that everyone who takes a CS course at Purdue must sign something of that nature before being given an account on UNIX.