Charging for a program almost necessarily entails concealing its source code and legally restricting how the software may be used, modified, and shared. For if the user pays for the source, and copyright or contract does not restrict what he may do with it in-hand, the demand for that product and the fact that copying is nearly cost-free together mean he has an incentive to distribute what he has acquired at a much lower price (possibly zero), as he does not need to recoup the capital its creator invested up front.

The non-technical minarchist might wonder what the problem with charging for software is. But there are two:

• the means of legally restricting user behavior amount to intellectual property law, and so are incompatible with a Rothbardian theory of property rights and contract law;
• technical users prefer having a program to having the experience of running a program.

The first should need no elaboration. However, exactly what a technically-competent user means by "having a program" probably does. The Free Software Foundation ties all this up in its eponymous term, defined thus:

… “Free software” means software that respects users' freedom and community. Roughly, it means that the users have the freedom to run, copy, distribute, study, change and improve the software. Thus, “free software” is a matter of liberty, not price. To understand the concept, you should think of “free” as in “free speech,” not as in “free beer.” We sometimes call it “libre software,” borrowing the French or Spanish word for “free” as in freedom, to show we do not mean the software is gratis.

You may have paid money to get copies of a free program, or you may have obtained copies at no charge. But regardless of how you got your copies, you always have the freedom to copy and change the software, even to sell copies.

We campaign for these freedoms because everyone deserves them. With these freedoms, the users (both individually and collectively) control the program and what it does for them. When users don't control the program, we call it a “nonfree” or “proprietary” program. The nonfree program controls the users, and the developer controls the program; this makes the program an instrument of unjust power. …

While not excluding the operation of the market process, this nevertheless presents a "market failure:" users who share these values of the Free Software Foundation desire software with the described liberties more than software without, but providing said liberties makes it difficult for the market to price it "correctly." A price makes it less valuable. This appears, on every level, to be a case where the neoclassical economist would identify a dead-weight loss, and prescribe some tailor-made state intervention.

But, possibly due to government's eternal and absolute technical incompetence, this intervention never materialized. And free software out-competes commercial software in every domain, save the desktop/laptop space, which is both older than the free software movement and has users generally desiring only the experience of running programs.

How?

The FSF's founder, Richard Stallman, ran into problems with commercial software at the MIT Artificial Intelligence Lab in the 70s. He characterizes its initial state as a "software-sharing community," where "[w]henever people from another university or a company wanted to port and use a program, we gladly let them. If you saw someone using an unfamiliar and interesting program, you could always ask to see the source code, so that you could read it, change it, or cannibalize parts of it to make a new program."¹ Commercializing interests, exploiting extensively intellectual property privileges, brought about the demise of this perceived utopia, tearing apart the society of the AI lab. Stallman responded by "building his own," from the ground up: an operating system and then a philosophical, social, and political movement, with the end of promoting the kind of environment surrounding software that was lost at MIT.

It is this GNU operating system, based exclusively on freely-licensed software, written by Stallman, the wider GNU project, and others entirely, whose derivatives dominate all serious computing in the current day. Not only did the non-state, entrepreneurial solution to the externality work, it out-competed its commercial alternative in a market that, if anything, had state actors hampering it (governments funding nonfree software startups, buying and using nonfree software, etc.). An estimated 54% of all computer devices shipped in 2015 running operating systems in approximate compliance with its principles (Android/Linux)². The software, rather than being written to be sold later, is funded ex ante or ex post facto, via crowdfunding and donations to individual maintainers/projects or to foundations that organize such development. These donations come from individual users, politically-motivated idealists, or corporations whose bottom line depends to some extent on the software in question. Lots of development is uncompensated, and are projects that a person develops primarily for personal use or interest and simply chooses to share for the benefit of others, at no cost to himself (indeed, often at a non-pecuniary benefit to himself: having users helps test the software, bazaar-style development means that development responsibilities may be delegated, and prominent open-source maintainership is an excellent recommendation to prospective employers).

A hard example for interventionists to chew on, indeed! It illustrates a common fallacy of those who would advocate state control: belief that entrepreneurs are somehow less capable of identifying and acting to correct externalities than the state.

Free software contains an additional lesson for libertarians: one in pragmatism. Stallman sought to establish a particular end in a legally-hostile system; ordinarily, there would be nothing preventing someone from copying, modifying, and redistributing the fruits of the project's labor under nonfree licenses, frustrating that end. So, he chose to exploit the legally-hostile system to prevent this practical harm to freedom, restricting users' freedom to make other users less free. Most of the GNU code proper is licensed under the GNU Public License, which contains so-called "copyleft" provisions, mandating that derivative works be licensed under the same terms. Copyleft uses copyright to destroy copyright, so to speak.

Often, libertarians fall into myopic fixations on the ideal free society, to the neglect of the reality that the fastest realistic path to it almost certainly will involve some liberty-maximizing stopgaps which, in isolation, would be unacceptable on libertarian terms³. Taking a page out of copyleft's book, adopting instead attitudes and policy prescriptions that celebrate and optimize for any increase in freedom, even if strictly incompatible with ideals, would go a long way towards securing libertarian ends.