Scapegoating free software’s failures

> Why can't a license restrict my freedom to build closed applications with free dev tools, in order to prevent the creation of closed software that likewise restrict others' freedom?

It can.

If just won't be a FOSS license by the traditionally understood definition of FOSS. For better or worse, the Free Software definition guarantees more freedom around end-user usage than it does around how you distribute software.

And maybe you think that's an arbitrary distinction. That's fine. You can build anything you want. You could build your own "Software Equality" or "Extended Copyleft" movement, or whatever you wanted to call it. Maybe it'll be even better than FOSS. The only thing you can't do is confuse people by calling your licenses something that they're not.

Nobody is upset that SSPL, API Copyleft, Parity, CAL, and Icepick exist. They're all fairly interesting licenses. I'm glad that they exist. Just stop trying to co-opt a movement and blur lines that have been clearly understood for years.

Large corporations are now stepping the unwritten meaning held by one portion who were previously satisfied with the overall outcome. Before, these people didn't care who owned the term as long as the goal was achieved. Now that the situation has changed these people want the phrase to mean what they always believed it meant. Even though you might not have been aware this difference even existed.

The meanings of words change with the progress of humanity. Nobody gets to hold on to a word like free just because they're comfortable with its current usage.

The current evolution of language like "free" is a political movement pushed by political actors; it's not based on an observation of how culture has evolved, but on a belief about how culture should evolve -- specifically how culture should evolve to benefit specific business interests.

In other words, Mongo didn't wake up one morning and discover that FOSS had changed, Mongo made a decision to try and change FOSS.

Of course, businesses and critics have the right to argue about FOSS definitions, and traditional FOSS advocates have a right to tell them to buzz off and stop stealing the positive associations that exist in no small part because we were dogmatic about definitions.

My position is that there are scenarios where dogmatism about words is genuinely important, and that sometimes, changing definitions should be resisted. The muddying of FOSS is no different from the muddying of words like "organic" or "humane" as labels for food products. Without a strong governing organization like FSF to prevent the dilution of these terms, the terms will quickly become meaningless and useless for end-users and consumers. FOSS may not be a purely prescriptive term, but that doesn't mean that every database product on the market should be able to manipulate it whenever they feel like they aren't making enough money.

I appreciate your frank admission of the distinction, but I have yet to hear a good defense of it that both rests on something other than "that's the way we said it is" and doesn't trap GPL or AGPL.

The understanding you describe exists, but I can't agree that it's broadly shared or accepted. If there's a tradition, it belongs to a very narrow tribe. That tribe isn't growing particularly fast precisely because its tenets are so esoteric and don't stand up to typical hacker interrogation by those not thoroughly marinated in the political history. That's the "byzantine, insular ideology of copyleft" I mentioned in the blog post.

> The only thing you can't do is purposefully confuse people by calling your licenses something that they're not.

I can call it open, and I do. Folks may hear both your view and mine, but the result isn't confusion. It's a choice.

There is no proprietary right in the terms "free software" or "open source". And there is no good reason for those presently ill served to pretend those ill serving them enjoy such a right. When a better mousetrap comes along, the designer of the last model doesn't get to pull the words "mouse trap" out of the newcomer's mouth. As in software, so in licenses: lead, follow, get out of the way, or get run over. When I follow my people to a license problem, I try my best to lead with solutions.

> Nobody is upset that SSPL, API Copyleft, Parity, CAL, and Icepick exist.

I'm glad to hear you say so, but I've received lots of feedback from people who are upset. And not just because "license proliferation". Have a look at the OSI license-review archives ... though I don't generally recommend that.

> The licenses you mention are not Free Software licenses because restricting what people do with their software, and not just how they share it, violates rule 0.

Making changes and building applications are things that people do with software, ways that people use software, in common parlance. The only folks I know who want to split that hair aim to defend FSF's free software definition. That definition was largely a post hoc rationalization designed to alleviate the cognitive dissonance between the increasingly libertarian, permissive bent taking over open source ideology via OSI/ESR and the overt prescriptivism of GPL. OSI revels in permissionlessness and an idealized, rules-free environment. GPL is full of rules.

It was politically expedient to draw a line around GPL and say "no rules, except these rules, because we need these people as allies". Not rigorous. Not otherwise coherent.

> Just stop trying to co-opt a movement and blur lines that have been clearly understood for years.

That movement hasn't moved in some time. The initiative is out of initiative. There is no widespread understanding of the details we're discussing. We are both rarefied wonks.

The movement you seem to think I'm coopting was itself coopted. Open source "won" by giving software freedom to developers who don't pass along to end-users in turn. I don't apologize for writing copyleft licenses that eschew the compromises of the earlier forms. Especially in light of their performance to date.