A useful rule of thumb when it comes to business models is to try to avoid things that depend upon other people acting in opposition to their short-term incentives. The whole concept of intellectual property, the idea that one can own and control the use of specific arrangements of data, is one such area. The ability to copy is ubiquitous, near free, and expanding into the physical world with the development of tools such as 3-D printers. When copying is less costly than purchasing from a source, people will copy. Such is human nature, and trying to go against this tide is a lot harder than accepting it for what it is and planning accordingly.

Clearly, as a practical matter in this age of computation, no-one can either own or control the use of specific arrangements of data. To do so would require the existence of a state and technologies of oppression so draconian and dystopian as to destroy all wealth and progress along the way to the goal of complete intellectual property enforcement. Such a state would make the worst excesses of 20th century communism seem mild by comparison. The end of that road is ubiquitous monitoring of every activity, state control of all computing machinery, and even the application of future neurotechnologies of thought and memory control to these ends. It is a path to be rejected outright, and rejected early. Yet the present state and its elite are committed wholeheartedly to this course, to the degree that today's technology allows, which might give some concern for the future.

To return to matters of today, here and now: if you produce a novel arrangement of data, whether a picture, a text, an algorithm, a theorem, a program, or anything else that can be usefully encoded and transferred as bits, you can only sell that arrangement once. After that sale, the arrangement is out in the world, freely available for anyone who wishes to obtain a copy. Any further revenues you receive after that first sale are actually donations, forms of charitable thanks. Yet individuals and organizations largely go through the motions of casting these donations as further sales, and build up a fanciful castle of concepts surrounding the idea that some transfer of ownership is taking place as a result. Another thing to avoid when it comes to business models is self-delusion, as it will lead to the undertaking of efforts that can only fail, based as they are on faulty axioms.

Release of formerly secret data after sufficient funds have been paid is known as the ransom model of development. Accepting donations following release is another form of patronage model. It is quite possible to mix and match for the same work of data arrangement. In both cases the number of patrons tends to be a small fraction of the number of people who make use of the data, but the approach to obtaining those patrons and the distribution and timing of amounts is quite different. My contention here is that there is actually no meaningful difference between these patronage models and the business of selling data as it is currently practiced in, for example, the software and writing worlds, other than the way it is presented, thought of, and talked about. But because the people doing the selling often think they are in the business of transferring property rights via contract, as opposed to soliciting freely given expressions of gratitude, they tend to do foolish and wasteful things as a result. One might as well be Canute set against the tide as to hope for the return of the same behavior as took place in past eras when the cost of copying was high.