Switch to MIT license

[omus]

I noticed this package uses the GPL-3 license. Is there any particular reason for this? Using GPL-3 may limit this package's adoption. Would you be opposed to switching the package to the MIT license?

[cjdoris]

I'm not necessarily opposed. Who do you think would not use it under GPL-3 but would use it under MIT?

[NightMachinery]

@omus commented on Jun 3, 2020, 9:38 PM GMT+4:30:

I noticed this package uses the GPL-3 license. Is there any particular reason for this? Using GPL-3 may limit this package's adoption. Would you be opposed to switching the package to the MIT license?

Freedom doesn't come for free. GPL gives free software developers a competitive edge.

[omus]

There's a good Discourse post on mixing MIT and GPL packages: https://discourse.julialang.org/t/gpl-and-virality/7715. In my use case GPL isn't an issue I just wanted to ensure the choice of the GPL license was intentional.

[cjdoris]

Thanks, that's a good article. Indeed the choice was intentional.

[iamed2]

JuliaDatabases/LibPQ.jl#209

[cjdoris]

@omus I don't use this package much any more, would you be interested in maintaining it? I'd be happy to re-licence it as MIT.

[iamed2]

He's away on leave for a while but I can commit on behalf of Invenia, which includes @omus and @fchorney.

[Byrth]

I detect a communication gap and shall insert myself to generate alert e-mails.

@cjdoris are you amenable to the proposal?

[cjdoris]

I'm happy to relicense this as MIT and add @omus and @fchorney as collaborators.

[mbravenboer]

How can we make this happen? It's a bit of a problem in the Julia ecosystem (via libPQ and Intervals)

In https://github.com/cjdoris/Infinity.jl/graphs/contributors I see that all contributors are in the discussion above, so presumably there are no other contributors who need to agree with the license change. Do any of the original contributors to this discussion make the change? If you need help with anything then the RelationalAI team can help too.

[cjdoris]

I have merged the PR to relicense this as MIT, thanks.

Note (see #25) that there is a new package Infinities.jl which combines ideas from this package and some other places. The intention is that everything is this package ends up there and we deprecate this one. It has types to represent unsigned/signed/complex infinity. The main thing from this package not there is InfExtendedReal etc.