For most people (as argued below) the internet’s single most important aspect is as a repository of algorithms, and so it is reasonably accurate to describe most people’s internet activity as “algorithm mining” (rather than “data mining”).

That’s why it is frustrating that there is so little literature on algorithm mining. E.g., the INSPEC database lists 29,045 articles on “data mining”, versus only eight on “algorithm mining” … and these eight are all false hits.

The concept of “algorithm mining” has both a narrow definition and a broad definition. Narrowly defined, “algorithm mining” might mean: “There is a complex system whose dynamics I want to understand/solve/predict/simulate; what mathematical algorithms, software tools, and databases are publicly available to help me?” And of course, the Internet is a uniquely powerful tool for algorithm mining in this narrow sense.

Broadly defined, “algorithm mining” might mean: “There is a complex system that I need to understand and engage; what narratives are available to help me understand it, and what communities of people embrace these narratives?”

In this broad sense (i.e., “social narratives are a class of algorithms”) doesn’t a pretty large fraction of internet activity center upon conceiving, constructing, and sharing algorithms, not data?

And therefore, aren’t the main challenges of Open Science centered upon algorithm sharing—including especially the politically thorny subclass of algorithm sharing, namely, building narrative consensus—rather than data sharing?

True, it is very common, and very convenient, to pretend that the Internet is all about data sharing … because all kinds of thorny political and social issues are sidestepped thereby … but that doesn’t mean this depiction is accurate!

As evidence, the algorithms—including narratives—provided by your links are (IMHO) far more interesting than the data.