Well on “other license” it should just warn that it _should be_ an open source license to be uploaded to hackage. That seems fine to me — its an informational message. Also note that we don’t accept packages with no license, just as we don’t accept AllRights licenses. So yes, the “no license” fact is true, but irrelevant. —g On March 7, 2015 at 12:50:39 AM, Carter Schonwald (carter.schonwald@gmail.com) wrote:

i'm very uncomfortable with the "warn on other-license" change. I think theres lots of valid reasons that someone may be using an amended license (eg BSD / MIT plus an explicit patent license grant) that strictly more open/free than any standard OSS license on the planet.

Edward Kmett raise the valid point on IRC that by current international treaties, authors no longer need to mark their works "All rights reserved" to protect their copy right, but rather that if no other license is specified, ANY work is by definition all rights reserved!

anyways, thats my 2cents for the evening

cheers-Carter

On Tue, Mar 3, 2015 at 1:56 PM, Gershom B wrote:

...

Thanks for this patch!

I've kicked off a discussion with hackage administrators and the haskell committee about the general approach we want to take to the license situation on hackage, and how to properly document our policies. It seems to me that merging this makes sense regardless, but I don't know what others may think?

Cheers, Gershom

On Tue, Mar 3, 2015 at 9:58 AM, Francesco Ariis wrote:

...

Dear Cabal developers, spurred by this discussion on haskell-cafe [1], I attach a small patch on licence warnings. It:

- reverts AllRightsReserved as PackageDistInexcusable, as it was before this commit [2]. Reading the comments in Check.hs, this datatype is for issues which "[are] OK in the author's environment but [are] almost certain to be a portability problems for other environments", which I think it is the case.

- adds a PackageDistSuspicious warning on OtherLicense. The text of the warning encourages the developer to choose from licences suggested by the OSI or FSF, if they don't want to use a licence recognised by cabal.

Thanks -Francesco

[1] http://mail.haskell.org/pipermail/haskell-cafe/2015-February/118411.html [2] https://github.com/haskell/cabal/commit/8d449ba3231445726272eac4dcf7b2b4a550...

_______________________________________________ cabal-devel mailing list cabal-devel@haskell.org http://mail.haskell.org/cgi-bin/mailman/listinfo/cabal-devel

_______________________________________________ cabal-devel mailing list cabal-devel@haskell.org http://mail.haskell.org/cgi-bin/mailman/listinfo/cabal-devel

there will never be an expressive enough licenses datatype. Law is complicated and fluid and changing. Period. On Sat, Mar 7, 2015 at 3:43 PM, Francesco Ariis wrote:

On Sat, Mar 07, 2015 at 12:50:19AM -0500, Carter Schonwald wrote:

...

i'm very uncomfortable with the "warn on other-license" change. I think theres lots of valid reasons that someone may be using an amended license (eg BSD / MIT plus an explicit patent license grant) that strictly more open/free than any standard OSS license on the planet.

I had a brief chat with dcoutts on freenode/#hackage, he informed me he would rather have the AllRightsReserved patch on hackage-server (and only in the public server branch) rather than cabal.

dcoutts also expressed similar objections on OtherLicense's warning (on the ground that dual licensing isn't supported by cabal yet, a a legitimate usage of OtherLicense).

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?").

Thanks for sharing your opinion! _______________________________________________ cabal-devel mailing list cabal-devel@haskell.org http://mail.haskell.org/cgi-bin/mailman/listinfo/cabal-devel

sure, my point is that there no universe where we wont HAVE to have an "otherlicense" option (ie, we MUST always have that escape hatch) likewise, i dont understand the dual license point. Could you explain a bit more? At the end of the day, auditing the licensing/intellectual property status of ones codebase/dependencies is subtle enough that i'm going to stay skeptical of any process that doesn't require human thought at the end of the line. For no other reason than its an insanely complicated topic. the law is about corner cases, "99%" is not a very good coverage. :) that said, i'm happy that someone is spending time thinking about this stuff, i'm just gonna try to push back on some anything which i think over reaches or veers into positing a legal opinion :) cheers! -Carter