Offering the example of Amazon suing Alexaholic (for remixing Alexa’s data), he tells us that APIs are not “a commons of goodies to be built on top of for fun and profit, like open source software.” Here are his “six basic truths of free APIs:”
- Free APIs are not a god-given right. Businesses offer them for their own self-interested reasons. If you build on top of the API but aren’t delivering the value for the business that provides the API, your use of the API will probably go away.
- If you build your own business on top of an API, you need a contractual relationship to ensure the service doesn’t get taken away from you. These generally cost money.
- If you find a way to get something from a site that isn’t explicitly offered as something for you to build on, your use of it will probably be fought unless you’re delivering value as in (1).
- The provider of your API will find it easier to implement services on top of their API than you will. Therefore you have to add something of your own that’s difficult to replicate, something beyond a simple UI tweak or a feature like “search”, so that the business that provides the API doesn’t simply compete with you when you look like you’re succeeding.
- For these reasons, free APIs are a very poor substitute for having the source and the data and thus owning and controlling every piece of your application.
- For these reasons, there’s no such thing as a free API if you’re looking to build a business.
Surely Torkington means free as in free beer APIs, as many of the problems he cites arise because the data and services are not free as in free speech. And this leads to two things I want us to be aware of in libraries: giving over our data to companies that lock it up behind licenses that restrict how it can be reused and remixed is dangerous; and we have an opportunity — some would say responsibility — to build out some of that information infrastructure and deliver free as in free speech APIs and data for all to use.