The GPL doesn't prevent the proprietary extensions and "pro" versions, it just prevents anybody but the original author from making those pro versions.

Therefore companies who want to make their money selling pro versions usually choose the GPL license. So in practice, we see more "commercial/pro" versions of GPL software than we do of MIT/BSD licensed software.

It prevents the "original" author from making those versions unless he or she requests a transfer of copyright for any outside code. If you have evidence of the second claim, I'd love to see it or see a link. (I mean this as non-confrontationally as possible; I don't have an opinion one way or the other on whether it's true, but now I'm curious.)

The standard example of this practice is MySQL. If you want some more, Google for "open core". Warning: you'll probably run into many heated arguments.

Re: the heated arguments, exactly the reason for my parenthetical comment earlier :). Thanks for the pointer.

> GPL [...] prevents anybody but the original author from making those pro versions.

This is only true as long as they accept no 3rd party contribution. In this case, accepting 3rd party modules is a crucial feature, so GPL+commercial wouldn't fly.

Yes, but in that case the "pro" versions, at least, are not parasitic on some other author's work

God forbid people want to get paid for their hard work and effort bringing something new into the world.

A copyleft license would not preclude this, as the Atom authors would still hold copyright, and could freely distribute a "Pro" version under a proprietary license.

A copyleft license would prevent someone else, who does not hold copyright on Atom, from doing that.

So you're happy that the original authors use copyleft+proprietary instead of allowing anyone else to release a fork (which would also be free software)? How's that better?

This is called "open core" strategy and few companies use it correctly because there is so much incentive to focus development on the "pro" version and let the free/open one stagnate (or lack basic features).

I wish they had it released under a less restrictive license like BSD or MIT.

http://en.wikipedia.org/wiki/Open_core

Well, it's also very difficult to justify contributing to a project that's open core because it means assigning your copyright to someone who stands to profit from your work while denying you the same right. There have been projects that have managed to still get contributions under this model [1], but they're probably the exception rather than the rule.

Of course, then when you don't get any contributions that only feeds the justification for focusing on the paid product, because clearly "the community doesn't care".

[1] mysql is the biggest and most obvious example, but it seems like a lot of tension built up over it and that's resulted in several forks existing now. Who presumably accept contributions without attribution to Oracle.

They did release it under MIT. It's right there in the first sentence.

I forgot that the first line of the GPL reads "YOU MAY NOT SELL THIS FOR PROFIT". It's a shame there are no billion dollar companies based around the GPL... /s

The /s indicates sarcasm.

Of course there is Google, Amazon, Red Hat, SUSE, even IBM uses a lot of GPL software...

All server software. No one makes money from FOSS desktop software.

You are, but the guys coding the software you use are not.

So unless you make the effort to contribute back, they will have some issues to pay the rent.

Firefox?

They get paid by Google.

Would Google have been able to get there if Linux and most open source software was licensed as GPLv3?

You would think people on HN would know better than to repeat such tired canards like this, but evidently not.

Copyleft licenses just make sure that people that want to get paid for their work (a legitimate venture) don't just piggyback on the collective effort of open source projects.

The GPL does not preclude this in the least.

Which is a problem because...?

Perhaps we will. I strongly doubt any of those will be even a slight success, unless released by the Atom team (which the copyleft license wouldn't affect).

Well, I'm glad for the closer to public domain MIT licensing.

I believe that the ability to sell it if you hit a home run leads to a lot of big ideas getting tried and the base hits making their way back into the open codebase.

But I think it is a matter of taste/personal preference. For me, copy middle feels more free.

That would be very sad.

It would preclude using so much code in extensions, for no particularly good reason.

JetBrains would disagree.