Jens Rehsack <rehsack%googlemail.com@localhost> writes: > Hubert Feyrer wrote: >> On Thu, 21 May 2009, Jens Rehsack wrote: >>> I'm currently updating devel/p5-DATA-UUID >>> (http://search.cpan.org/dist/Data-UUID/) and there is as LICENSE simply >>> named 'BSD' with the following text: >> ... >>> Which license is meant? The original BSD license or the modified one? >> >> Why do you think it's a BSD license at all - just by the name of the >> file? > > No, but by the statement in META.yml: > 'license: BSD' there is explicit text. You have to diff the text with existing files to see if they match. It makes no sense to look at the operative properties and say it's sort of like a BSD license. In this case, it's not a BSD license. There's a notion that permission is granted only to those who acknowledge lack of warranty or something. That's not intrinsically unreasonable, but it's a huge legal change, from bare license to an attempt at a contract license (which may or may not make sense). Just because someone else put "license: BSD" in META.yml (whatever that is) doesn't make it so. That could well be someone else's attempt to sort licenses into two bins, or it could be wrong. Again, I think the benefits from tagging free licenses are minimal, and cases like this I'd be happy to leave a # License is more or less BSDish, but not exactly. # LICENSE= in pkgsrc and let it be until later. It's also reasonable to ask upstream to clarify the license to *exactly* the text of an OSI/FSF-approved license, but that can only be done with permission from all copyright holders so may not be all that likely. I think what this is all pointing out is that there is a lot of pain From pkgsrc being on the leading edge of licensing purity/clarity.
Description: PGP signature