
On Sun, Mar 08, 2015 at 01:19:34PM -0400, Carter Schonwald wrote:
On Sat, Mar 7, 2015 at 3:43 PM, Francesco Ariis
wrote: My view is that, with an expressive enough License datatype which covers an ample portion of usages, the warning could still be pragmatically useful ("do you really have a reason to draft a new document when there is probably something tried and tested out there which could do for your case?").
there will never be an expressive enough licenses datatype. Law is complicated and fluid and changing. Period.
Well, I ran a little test on the index of packages [1], to check the most popular licences there and see how widespread is the use of OtherLicense. BSD3 5007 MIT 976 GPL 460 OtherLicense 307 GPL-3 286 PublicDomain 199 LGPL 145 GPL-2 81 Apache-2.0 53 LGPL-3 51 LGPL-2.1 49 parse-error 43 none 36 BSD2 23 AGPL-3 21 BSD3 8 BSD4 5 OtherLicense 3 Apache License, Version 2.0 3 LGPL-2 2 <Misc> 13 Parsing was extremely crude, but enough to conclude that OtherLicense amounts to less than 4% of the total amount of packages (7771). If we find a way to deal with dual licences and add some missing licences to Cabal (e.g. Artistic License 2), the Licence datatype will cover 99%+ of usage, which is expressive enough in my opinion (and it's not we cannot add more stuff as new licenses pop up). [1] https://hackage.haskell.org/packages/index.tar.gz