Subject: RE: ACCEPTABLE_LICENSES
To: None <tech-pkg@NetBSD.ORG>
From: Greg A. Woods <email@example.com>
Date: 06/10/1999 14:40:35
[ On Thursday, June 10, 1999 at 06:51:42 (+0100), Alistair Crooks wrote: ]
> Subject: RE: ACCEPTABLE_LICENSES
> You have missed the point.
> Acceptable licences does not govern whether or not the distfile can be put
> on a CDROM or stored on a mirror or other server.
No, that's exactly my point! It's used for too many things and confuses
*far* too many issues.
> Acceptable licences is a way for a user on a NetBSD machine to specify that
> they do not want commercial, shareware or other "non-open-source" types of
> software installed on their machine. This means that, if I want to be abide
> by the licence terms which people put on their software, but am unwilling to
> pay for it, for example, then any bulk build in pkgsrc will miss that
> software because I'm not prepared to cough up the dough for it. You may
> think said licences are untenable, or are malodourous pieces of junk, or
> only applicable to people in the 48 states - I don't care. We needed a way
> to provide users, on a site-basis, with a way of avoiding shareware, if
> necessary, and I'm quite pleased with the way it works.
The problem is that no one group of people can even hint at offering a
proper classification system for third-party software, especially
without being far more meticulous, impartial, and discriminating in
defining the terms they use and setting out standards for their usage.
The things you mention are so totally subjective and so totally removed
from the real impact of the law (at least in the legal jurisdiction
where I reside) that they are misleading and harmful. Perhaps if
"ACCEPTABLE_LICENSES" were to simply match against impartial tags such
etc., then perhaps I could support it's use for the purposes you
describe. However even to go that far would require some very careful
analysis of the actual copyright licenses in various bits of software --
is there an international intellectual property lawyer on staff to do
this analysis and make legal recommendations?!?!?!?
In the end it wouldn't really achieve anything anyway.
I believe there are only two valid concerns w.r.t. copyright law that
pkgsrc should deal with: is the distfile anonymously available? and, is
it re-distributable? The only reason for making a distinction between
these two is because there's lots of freely available software that
cannot be redistributed (usually only because the author(s) didn't
understand the copyright terms they were writing), and one of the
features of pkgsrc is that it can be used to create a collection of
software that can easily be made available for redistribution.
If the distfile is available by anonymous means then the copyright
license CAN NOT restrict how one runs the software or for what purposes
it is run (at least not in most western legal jurisdictions that I'm
If the distfile is not anonymously available (eg. the 128-bit Netscape
binary) then it is the user's sole resposibility to understand and
adhere to the restrictions that apply to it before obtaining a copy.
Users are *entirely* responsible for their *own* actions. The less the
pkgsrc system tries to make decisions for them then the less chance
there is that TNF can get into trouble, and the less chance the users
will get themselves into trouble.
Other concerns w.r.t. to non-copyright issues, such as patent protection
and export controls are much more sensitive to the locale of the user,
and are much more difficult to clearly and impartially classify in a
centralised system like pkgsrc. I think that about the best pkgsrc can
hope to do is to list packages that contain crytpo and which must be
downloaded from the USA and those which contain crypto but are available
from non-USA sources (and hopefully which haven't been illegally
exported from the USA, Canada, etc.!).
> As far as I'm aware FreeBSD have not used the acceptable licences stuff.
> That's their problem, not ours.
Actually they started with ACCEPTABLE_LICENSES (if I remember correctly)
and then eventually saw the fallacy of their ways (i.e. what I'm trying
to point out, again) and switched to a system that was applicable in the
only place it made sense to apply it. I.e. they don't have a problem
Greg A. Woods
+1 416 218-0098 VE3TCP <firstname.lastname@example.org> <robohack!woods>
Planix, Inc. <email@example.com>; Secrets of the Weird <firstname.lastname@example.org>