Subject: Re: copyright questions
To: Wolfgang Solfrank <firstname.lastname@example.org>
From: Chris G. Demetriou <email@example.com>
Date: 06/13/1997 14:24:11
> > I don't specifically remember what went on with 1.0, and don't have
> > easy source tree access to check, right now.
> Well, I still have available the discussion between you and myself about
> the Legal Mumbo-Jumbo in 1.0. The discussion took place a few days after the
> release of 1.0 (because I didn't look into the docs before), and you
> specifically mentioned that there wasn't a requirement to include all
> contributors in that file, because the release announcement wasn't
> advertising material and it didn't mention all the parts of NetBSD
> specifically. I could argue that that was a violation of the spirit of
> the copyright message.
Yes, that's certainly a violation of the spirit, if not the letter.
(Even if the release announcement was advertising, if it didn't
mention specific features or use, under the language of the standard
berkeley-style license it still wouldn't have had to include the
I think part of the problem is what I mentioned: It's a ... huge PITA
to retroactively search through the source tree and find all of the
attribution strings, unless that's legally required. It needs to be
done well once, and then that documentation needs to be kept up to
date as things are checked into the source tree. Unfortunately, doing
the first pass is a huge amount of work, and doing the ongoing
maintenance is a large amount of work as well.
> > (1) I made a good faith effort to include all of the appropriate
> > attribution strings (i.e. "This product contains software written by
> > ...") in the release documentation. If I missed a few, that's
> > unfortunate. However, I really did try to include all of them.
> I don't think so. The release documentation included only a subset of
> the contributors.
That depends on how you define "appropriate."
If you define appropriate by the "letter of the license," then I made
a good faith effort to do that, yes.
If you define appropriate by the "spirit of the license/project," then
you should probably be more severe than that: the release
documentation included only a subset of the contributors who actually
had copyrighted source in the source tree with an attribution string
Like i said before, in a perfect world, everybody who has done
anything for the project should be credited, and that credit should
include some description of their contributions. However, it's not
practical for a release engineer to do that at release time; there has
to be some prior preparation.
Perhaps I'm (partially) at fault for not being more careful about
making sure that such a list was created to begin with. Perhaps I'm
(partially) at fault for not requiring people to donate code, rather
than simply having them put it in the source tree with their own
copyright and license terms. Indeed, if I were going back and
starting things over (or, if I were still one of the people in charge
of the project) then those are things that I would seriously consider
In any case, looking back and thinking of how I or we might have done
things differently doesn't have much immediate positive effect. If I
offended you then, I apologize.
> I don't want to sound rude, and I'm happy (more or less) with your new
> license, but it looks to me as if your attitude regarding this has changed
> a bit since then.
I could believe that. It's been a long time (mmm, long release
cycles 8-), and I've certainly got a lot more experience now.
To some degree, it's also a case of the side of the fence i am or was
sitting on. I don't think I was ever _happy_ about not giving people
as much credit as they wanted or deserved. However, as a release
engineer, it was pretty much impossible to justify or take the time to
do the "right thing" (where "right" == by the spirit of the license,
rather than the letter) when the time to do the "right thing" was so
enormous. Certainly, others pushed me to do the "right thing," but
requests and demands don't necessarily confer time to do the work.
Even now, release documents don't typically include more than the
minimal legally-required set of attribution notices. (Look in the 1.2
release notes for e.g. the i386. You're listed as "supporting cast,"
but you still don't have any attribution string included for your
> PS: I can make available your original answer if you give permission.
I don't think so. 8-)
However, could you send a copy of it to me? I'd like to see it, and I
probably don't have a copy archived...