Re: MMIX license OK for main?
Excuse me for reopening a month-old thread, but browsing the archives I
thought I had something of interest (email from Knuth, no less! albeit a bit
old) to add to this one.
Mike Goldman wrote:
>"J.H.M. Dassen (Ray)" wrote:
>
>> On Thu, Nov 11, 1999 at 21:37:13 -0800, Bruce Perens wrote:
>> > There's got to be someone at Stanford who can get to him.
>>
>> See http://www-cs-faculty.stanford.edu/~knuth/email.html for contact
>> details.
>
>Yes, but to wait 3 months for a response to a snail mail request certainly won't
>resolve the issue in time for Potato.
>
>If we cannot get a clarification of the license before early December, I'll
>repackage MMIX as source only, plus a build script, ala Pine.
I'll quote the copyright notice being discussed, for those who have forgotten
> copyright 1999 Donald E. Knuth
>
> This file may be freely copied and distributed, provided that
> no changes whatsoever are made. All users are asked to help keep
> the MMIXware files consistent and ``uncorrupted,''
> identical everywhere in the world. Changes are permissible only
> if the modified file is given a new name, different from the names of
> existing files in the MMIXware package,
> and only if the modified file is clearly identified
> as not being part of that package.
> (The CWEB system has a ``change file'' facility by
> which users can easily make minor alterations without modifying
> the master source files in any way. Everybody is supposed to use
> change files instead of changing the files.)
> The author has tried his best to produce correct and useful programs,
> in order to help promote computer science research,
> but no warranty of any kind should be assumed.
I'm not at all knowledgeable about Debian or its Policy, but I have had a
brief exchange with Don Knuth back in 1994 about CWEB, which has a licence
similar to the one of MMIXware; indeed I think the latter has evolved from the
former. The current copyright for CWEB is
% The CWEB programs by Silvio Levy are based on programs by D. E. Knuth.
% They are distributed WITHOUT ANY WARRANTY, express or implied.
% Copyright (C) 1987,1990,1993 Silvio Levy and Donald E. Knuth
% Permission is granted to make and distribute verbatim copies of this
% document provided that the copyright notice and this permission notice
% are preserved on all copies.
% Permission is granted to copy and distribute modified versions of this
% document under the conditions for verbatim copying, provided that the
% entire resulting derived work is given a different name and distributed
% under the terms of a permission notice identical to this one.
I had produced a modified version of CWEB, in fact based on something forked
off several years earlier by somebody else, and had decided to make that
version public without changing the name (this was foolish, though I think not
illegal, but not the main point here). The copyright notice at the time of
forking was slightly different from the one just cited:
% Copyright (C) 1987,1990 Silvio Levy and Donald E. Knuth
% Permission is granted to make and distribute verbatim copies of this
% document provided that the copyright notice and this permission notice
% are preserved on all copies.
% Permission is granted to copy and distribute modified versions of this
% document under the conditions for verbatim copying, provided that the
% entire resulting derived work is distributed under the terms of a
% permission notice identical to this one.
You see, Knuth added the clause about giving a different name to modified
version; I presume this was a reaction to variants of CWEB appearing (and
apparently in violation of the (old) copyright notice itself!).
So let me now quote the reaction Knuth gave me, after I had contacted his
co-author Silvio Levy. I'll give it in full, straight from the horse's mouth
> Date: Tue, 8 Feb 94 23:36:58 -0800
> From: Phyllis Winkler <winkler@CS.Stanford.EDU>
> To: maavl@cwi.nl
> Cc: levy@geom.umn.edu
> Subject: A note from Don Knuth regarding "CWEB 3.x"
> Reply-To: winkler@CS.Stanford.EDU
>
> I was quite disturbed by your announcement, and also surprised by your
> remark that CWEB didn't seem to be changing... when I was in Germany
> last summer, I found many centers would have every change to CWEB
> installed on servers for the rest of Germany [and I thought the rest
> of Europe] within a few hours of the changes made on labrea at Stanford.
> CWEB went through 2.2, 2.3, 2.4, ... 2.9, 3.0--, 3.0-, and 3.0 before
> hitting 3.1!
>
> Silvio and I don't want anybody else to assume responsibility for any
> system called CWEB. Whether you personally like to have another name for
> your system or not, you mustn't call it by the name we are using, since
> that can only be confusing. Since we are assuming the responsibility
> for maintenance, we don't want anybody else screwing around with the
> sources either. All suggestions for changes have traditionally been
> filtered through Silvio and me, and we check each other that they have
> been correctly made.
>
> However, we are not averse to progress! We realize that CWEB is not by
> any means the best possible system for literate programming. We just
> don't want the endless confusion that would exist if the system became
> divergent. We also believe that the system as it stands, warts and all,
> is plenty good for all of our future needs. We are well aware that any
> complex system is imperfect and can be endlessly enhanced, but after
> awhile it is better to have convergence. So we worked hard to get CWEB
> into a stable state where we could say, finally: Enough! And we published
> the source code in a book [Addison-Wesley, December 1993], after which we
> have decided not to make any more changes unless some catastrophic error is
> reported. (Even then, we may decide to declare the bug a "feature".)
>
> Thus, we ask you most emphatically to choose another name for any enhancements
> you feel compelled to make for your purposes. You can, of course, distribute
> your system to whoever you want. To minimize confusion it would certainly
> be best if you could be 100% upward compatible with CWEB. Since we aren't
> changing CWEB any more, you don't have to worry about a moving target;
> compatibility today means compatibility tomorrow.
>
> A change to the syntax, like making @<Section names@> statements instead
> of expressions, is incompatible and therefore a disaster in my opinion.
> A "compatibility mode" is one way out; notice that the program now is
> already set up to understand many command line options presently ignored.
> Users can set aliases so that their preferred defaults are automatically
> supplied on the command line.
>
> When future hypertext environments are built for literate programming, it's
> reasonable to assume that the builders will try to make their systems work
> with the tens of thousands of CWEB programs that will be in networks by then;
> that's another reason to stay as compatible as possible. There may also be
> hundreds of thousands of programs written in your system, whatever it's
> called. But I am not going to keep twiddling with tools when I want to
> use them to write hundreds of programs for Volume 4; and I'm sure Silvio
> feels the same. Look, Wirth never called his languages FORTRAN.
>
> I do believe you should respect our wishes in these matters, because we put a
> great deal of effort into getting a stable system that we could both live with
> and assume ultimate responsibility for. When you take responsibility for your
> own system, you may wish to spend the rest of your life enhancing it, or you
> may come to the same position we are now in: where you say THIS IS WHAT I WANT
> TO USE FOR THE NEXT N YEARS AND I DON'T CARE WHAT IMPROVEMENTS ANYBODY ELSE
> MAKES I JUST WANT TO GET ON WITH MY REAL WORK AND IT'S FINE IF OTHER PEOPLE
> WISH TO MAKE SYSTEMS THEY LIKE BETTER AS LONG AS I DON'T HAVE TO BOTHER TO
> LEARN THEM.
>
I guess this sums up Knuth's opinion nicely, and I have no reason to doubt
that he feels the same way about MMIXware, or for that matter, about the
Stanford GraphBase. I leave it to you experts to find out what this means for
the distribution licences.
--
Marc van Leeuwen
Universite de Poitiers
http://wwwmathlabo.univ-poitiers.fr/~maavl/
Actually I'm extremely glad to see the continuing development of
languages, not only because programming languages are getting
better and better in important ways, but also because such work
soaks up a lot of people's energy---therefore computer scientists
don't write papers that I would otherwise have to read, and I can
get my books finished a lot sooner.
-- Donald E. Knuth
Reply to: