Re: Concerning uniliateral decision making with respect to the coreutils package

Simon Josefsson <[email protected]> Tue, 14 Jul 2026 12:06:57 +0200
Newsgroups	gmane.linux.debian.devel.general
Message-ID	<[email protected]>
"Theodore Tso" <[email protected]> writes:

> On Mon, Jul 13, 2026 at 10:30:44PM -0500, Simon Josefsson wrote:
>> My point earlier in the thread was that by replacing (strongly) copyleft
>> software by non-copyleft software (and I include weak-copyleft in the
>> non-copyleft category) in Debian there is a risk the technical
>> excellence of Debian will be used to subjugate users, and that
>> eventually this could erode the foundation for why Debian is technically
>> excellent in the first place.
>
> The way to keep Debian "technically excellent" is use technically
> excellent components, regardless of whether they are licensed using a
> strong-copyleft, weak-copyleft, or a permissive license.  I believe
> that choosing technically inferior software because it has a license
> that we might find preferable is ultimately self-defeating.

Let's say in 10 years that companies have sponsored BSD/Expat-rewrites
(possibly LLM-assisted) of essentially all common GPL'd tools, and has
showed multi-year sustained QA and release process of those projects.
Compare OpenSSL, Clang or UUtils as inspirational projects.

At the same time, development of GNU coreutils, sed, tar, gzip etc have
slowed down and are in maintainance mode.

It would then be easy to make the argument that those rewrites are
"technically excellent".

Do you think there could be some reason beyond "technical excellence"
that would make us want to keep using the (strong) copyleft projects?

What's left of the spirit of the DSG and DFSG if copyleft software is
replaced?

> We could talk about some of your other comparisons: Seqoia vs GnuPG,
> Clang vs GCC, but I note that you omitted one of the major counter
> examples to your thesis --- namely, the Linux Kernel vs FreeBSD.
> That's an example where the strong copyright licensed component has
> been more popular, and more features, than the permissive-licensed
> kernel.  So I don't think the death of strong copyright is inevitable.

Yeah, I think by using the Linux kernel my argument becomes more
subtle/complex: then I rely on the enforceable property of strong
copyleft.

Another way to weaken the ideas behind strong copylefts is to use GPLv2
but in parallel say "I won't bother suing anyone violating the license,
and think it is a waste of time".  That's pretty similar to a
non-copyleft situation.  We need to enforce our license terms for them
to be useful.  We'll see how that develops.

> I used to use the proprietary software Parallels to run Debian on
> MacOS, but I'm now using the Apache licensed Container CLI.  This may
> not be your preferred strong copyright license, but it's an example of
> moving from Proprietary -> FOSS, which hopefully you'll agree is
> preferable.

Yes.  I find the objections to using proprietary software a bit
idealistic.  Sounds like your use of Parallels helped improve Debian.
We wouldn't have a lot of free software if people didn't start using
computers running proprietary software.  People should be able to run
what they want.  Using Windows, macOS or iOS (or Debian) is a great way
to get inspiration for doing better things.

This user freedom includes using free software without any proprietary
blobs too, if that's someone's desire.  That's why I dislike not giving
Debian's users that possibility, with the new non-DFSG installer images.

If we show this is possible, maybe someone eventually won't find any
need to use proprietary software, but it should be a user choice.

> It's not a matter of popularity, but using the best possible tool for
> the job.  And if you want to see copyleft software succeed, the best
> way to do that is contribute to make it better.  Not to try to force
> or gatekeep users to using softare that isn't best suited for them
> just because of the license --- that way only lies the death of our
> credibility.

I wish copyleft software will be sustainable but I see big risks if we
don't make an active choice to actually support the software that we
believe will promote user freedom in the long term.

/Simon
signature.asc (application/pgp-signature, 1.2 KB)