Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/18/84; site Shasta.ARPA
Path: utzoo!watmath!clyde!burl!ulysses!allegra!mit-eddie!genrad!decvax!decwrl!Shasta!linton
From: lin...@Shasta.ARPA
Newsgroups: net.emacs
Subject: dbx distribution revisited
Message-ID: <6447@Shasta.ARPA>
Date: Mon, 24-Jun-85 23:31:01 EDT
Article-I.D.: Shasta.6447
Posted: Mon Jun 24 23:31:01 1985
Date-Received: Wed, 26-Jun-85 06:32:58 EDT
References: <2355@topaz.ARPA>
Organization: Stanford University
Lines: 43
I should clarify my response to Stallman's message about dbx distribution.
I should also expand on my perspective on the legal domain of dbx source.
I have written dbx almost entirely from scratch. There were a couple
of places where I looked at adb source to figure how things were done
(e.g., read a coredump), and one place (instruction disassembling)
where I started with several hundred lines of adb code and modified them
to work with dbx. The modifications included a substantial style change, as
well as a few changes to type names and control flow. The resulting source
hardly resembles the adb code.
Since I neither understand nor particularly care about the legal issues
involved, I have avoided worrying about whether this history makes dbx
dependent on an AT&T license. My opinion is that my opinion on this issue
doesn't matter, and if you care you should consult a lawyer.
Now, other people are interested in whether dbx is AT&T dependent, and
when I tell them the history they come to their own conclusions. I also
tell them what I said above, that my opinion is they should get a lawyer's
opinion (unless they decide it is AT&T dependent, which is certainly safe).
When I told Stallman, he interpreted it to mean that dbx was not AT&T
dependent, and that it could thus be freely distributed. That is certainly
his prerogative, and I did not mean to say that he is wrong. In fact,
I would be glad if he is right.
The reason I objected to his first message is that I thought he made it
sound like I wrote dbx completely independent of AT&T code and that
I had said that dbx was public domain. I know that he does not
believe either of these, and he meant to state his interpretation
of the facts that I relayed to him. Unfortunately, the difference
between <what I said> and <what he interpreted from what I said>
was not clear.
Finally, I want to emphasize that I did NOT mean to say that Stallman
is wrong in his interpretations, or that he purposely misrepresented
my opinion. I understand how he reached his conclusions, and they
are certainly reasonable. Do not take my refusal to support them
as disagreement -- I simply don't have the expertise or interest to
come to my own conclusions.
Mark
Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/18/84; site masscomp.UUCP
Path: utzoo!watmath!clyde!bonnie!masscomp!z
From: z...@masscomp.UUCP (Steve Zimmerman)
Newsgroups: net.emacs
Subject: Re: dbx distribution revisited
Message-ID: <730@masscomp.UUCP>
Date: Wed, 26-Jun-85 09:50:45 EDT
Article-I.D.: masscomp.730
Posted: Wed Jun 26 09:50:45 1985
Date-Received: Thu, 27-Jun-85 05:47:34 EDT
References: <2355@topaz.ARPA> <6447@Shasta.ARPA>
Organization: Masscomp - Westford, MA
Lines: 31
The legal issues involved here certainly are somewhat messy. In the
case of AT&T code, even greater care must be taken compared to something
like Gosling's code, since the AT&T code is protected by trade secret
protection, which is considerably broader in scope than simple copyright
protection. Under trade secret protection, algorithms and ideas as well
as code are protected. To definitively say how that would relate to a
case such as dbx is hard to do without expert legal advice, but there
are some guidelines. For example, Joe Chapman of CCA Uniworks stated
the following in a recent message to this newsgroup:
Beginning with CCA EMACS version 162.36z, CCA EMACS no longer
contained any of the code from Mr. Montgomery's EMACS, or any
methods or concepts which would be known only by programmers
familiar with BTL EMACS of any version.
This is essentially the same wording used by AT&T in a letter to CCA at
the time CCA and AT&T resolved the situation regarding CCA EMACS and
Warren Montgomery's code, and is a good indication of what AT&T
considers trade secret protection to mean.
Which side of the line does dbx fall on, and how much risk is there in
using it for sites without source licenses? I certainly can't offer a
definitive answer here. However, I do know that during my time as
project manager for CCA EMACS I dealt quite a bit with the CCA lawyers,
and they would never have let me put something like dbx into the CCA
EMACS distribution. So I would suggest that individuals either check
with their own company attorneys on the matter before using this code,
or else that a qualified attorney post an opinion to the net.
Steve Zimmerman
Masscomp