[reply restricted to list] At 2022-02-28T13:34:09+0000, Jonathan Wakely via austin-group-l at The Open Group wrote: > On Mon, 28 Feb 2022 at 13:03, Robert Elz <k...@munnari.oz.au> wrote: > > > > Date: Mon, 28 Feb 2022 10:28:20 +0000 > > From: Jonathan Wakely <jwak...@redhat.com> > > Message-ID: > > <CACb0b4k8H78coiC24S-wpna3GaEO=vy37opu_1-vf_ybugg...@mail.gmail.com> > > > > | Nothing in any GNU licence prevents reading code. > > > > Not explicitly, no. But if I read some code, and then write > > something similar, how would I ever prove I had not copied? > > If I did, even subconsciiusky, and distributed my code, it > > would be required to be GNU licensed wouldn't it? That is > > something I will never do. I give away code I write unrestricted > > (free). Not encunbered ir restricted. > > How is that different from any code that isn't in the public domain? > It's copyrighted either way, and you are bound by the terms of the > licence.
Indeed. I worked for several years in FLOSS compliance at a Big Company(tm) with hundreds of distinct products, and the third-party software copyright disclosure PDFs for projects that attempted to minimize the footprint of copyright could run to dozens or even over a hundred pages, driven mostly by BSD- and Apache-style licenses. (Those "NOTICE" files add up if you pay attention to them.) Evidence of this phenomenon is not hard to come by--talk to anybody familiar with the field, or check out the fine print of the documentation for your smart TV or Android phone. (Frequently, it is offered with little or no explanation for its presence.) > > Hence, and especially here as the starting point was that I > > might add code to the BSD realpath utility to make it more > > like the coreutils version, and hence make it easier to > > standardise, I simply cannot look at the GNU code or I would > > not be able to distribute a modified BSD version under the > > BSD licence. > > This is simply wrong. You can look at it. If you're unable to > reimplement it without making it a derived work, then either you > really *are* copying it, or it's not really copyrightable in the first > place (e.g. it simply does something so trivial that there's only one > way to do it, and nobody is going to be able to argue that you > "copied" the original just because you did the same thing). Yes. Copyright is grounded on the presence of original expression and the practice of authorship (as opposed to scribal activities like transcription)[1]. Where there is no room for that, it is difficult for copyright to here. I do not say "impossible" because a sufficiently wealthy and fortunate party can always achieve cockeyed results, whether in the courts or through intellectual property treaties. > What in the GPL means your eyeballs are tainted, This query cuts to the chase. What we're dealing with here is, I submit, not really a matter of legal restriction but a puritanical stance that is scandalized by copyleft, and locates virtue in tactically restricting access to source code. Whether this is done in conscious emulation of AT&T Unix source license management practices starting in the mid-1970s, or because such restriction is considered inherently virtuous seems to depend on the person. The clubhouse/"core team" approach to source project management reproduces this dynamic with respect to write privileges, and played out in exemplary ways with the NetBSD/OpenBSD split and the XFree86's self-exile into irrelevance. > or what in the BSD licence says you don't need to conform to the > licence if you copy code? The entire USL v. BSDI settlement turned on this very point[2]. 4.4BSD was able to rise like a phoenix from the ashes of the pyre USL had built for it because USL and/or its predecessors-in-interest took a lax attitude to the terms of the UC Regents' copyright notice, and duplicated code without copying the requisite notices. Perhaps the fact that this information came to the awareness of the general public only concomitantly with the SCO vs. IBM lawsuit over the Linux kernel (itself surely a point of amusement to some Linux/GPL detractors) explains why so many BSD partisans today remain ignorant of it. > Please stop calling it "stupid" based on an apparent misunderstanding > of how copyright and licensing works. This discussion doesn't belong > here, but wouldn't be necessary if you'd avoid the inflammatory > language. In other areas, this stance is familiar as the "politics of resentment"[3]. It would be unfortunate if the (further) standardization of readlink/realpath commands by POSIX were frustrated, but especially so if said failure were falsely attributed to the application of copyleft licensing to a particular implementation. (I remember when true(1) and false(1) were shell scripts that proclaimed themselves to be trade secrets--"unpublished proprietary source code of AT&T"--in full caps.) To avoid that outcome, I volunteer to prepare a human-readable English[1] description of the behavior of the GNU coreutils readlink and realpath commands' handling of the '-e' option. Or someone else can do it. Either will dispose of this spurious, and splenetic, objection. Regards, Branden [1] https://www.supremecourt.gov/opinions/20pdf/18-956_d18f.pdf See particularly page 12. [2] https://www.itreview.org/documents/2005/0623.html [3] https://verfassungsblog.de/understanding-the-politics-of-resentment/
signature.asc
Description: PGP signature
