On 12/9/23 05:26, Daphne Preston-Kendal wrote: > Speaking personally, I have very little interest in debates about choice of specific license, but as WG2 chair I wanted to add this: > > For us it’s important that the licence attached to sample implementations allows the SRFI sample implementation to be used as a base for a sample implementation for the corresponding R7RS Large version of a library. > This indeed seems like a very important consideration. > We haven’t actually discussed yet what licence the R7RS Large sample implementations will appear under, but I suspect it will be a maximally permissive and GPL-compatible licence because they should be usable by Schemes with all sorts of different licensing arrangements. (The ideal solution from that point of view might actually be CC0/public domain, but this is flatly impossible unless we actually produce new sample implementations of everything *or* persuade the relevant SRFI sample implementation authors to apply this to their work.) > I was (relatively) recently surprised to learn that CC0-1.0 was not accepted by the OSI as an Open Source license (though it was also not rejected, but rather withdrawn from consideration), and Fedora recently changed their policy to forbid CC0-1.0. As I understand the argument, while many approved licenses don't explicitly discuss patents, they may provide an implied patent license, whereas CC-0 includes language making explicit that it is only a copyright license (well, and "sui generis database rights"). There's some relatively recent discussion at <https://github.com/fsfe/reuse-docs/issues/62>. This has not stopped me from using CC0-1.0 myself for files I do not believe to be copyrightable, and certainly not patentable, but it would probably create unnecessary problems for R7RS Large. The cleanest solution would be for Creative Commons to release either an alternate version of CC0 or a supplemental patent grant that could be combined with CC0-1.0. > > Text within the SRFI documents may also be adopted into the report. The licensing situation of the Scheme reports is not ideal, but it’s not practical to fix it after all these years: <https://codeberg.org/scheme/r7rs/src/branch/main/LICENCE.txt> > I’m less concerned about this as there are various solutions (including simply rewriting the specification text for any adopted procedures/syntax/etc.) we can adopt. > In the process of adding license metadata to Racket packages, Sam Tobin-Hochstadt and I got this recognized by SPDX as the SchemeReport license: see <https://spdx.org/licenses/SchemeReport.html> and discussion at <https://github.com/spdx/license-list-XML/issues/1340>. I had also almost finished a request to the OSI to approve it, but I guess I never actually sent it: I'll try to do that soon. Philip
