Brussels / 3 & 4 February 2018


The Future of Copyleft: Data and Theory

Using (copylefted!) data from the 32 repositories and 2.5M projects covered by, we'll survey the state of copyleft. This will include the growth of AGPL, the reciprocal scope of the GPL, and what stacks copyleft has been most successful in. We'll then use the data to inform a look at theory and discuss where copyleft might be going: where is copyleft's success? where and how is that relevant in the modern software landscape? what directions might future copyleft licenses take?

General note: scheduling permitting (they're also running a dev room), I may be joined by Andrew Nesbitt or Ben Nickolls, co-founders of

Many of the various assessments of license "popularity" have been skewed either by using a very limited data set (e.g., Debian, Fedora, which cover ~1/100th of FOSS) or by being proprietary/unreproducible (Black Duck, etc.) In this talk I'll discuss another open data source by diving into the licensing data from, which we believe to be the largest open repository of information about packaged FOSS. Because it includes dependency data, it can tell us not only about numeric usage, but also about relative importance and position in the stack of various licenses; and since it has a notion of "stack" (tied to repositories) it can inform some of our intuitions about how license usage varies by stack. We'll use this to assess

(The biggest shortcoming of the data is that, because there are no "repos" per se, it does not cover C/C++/core operating system components; I'll also discuss this in the talk.)

After discussing the current state of copyleft using data, we'll discuss what the future of copyleft might look like. This discussion will be informed by the data, but not limited to the data. Among other things, I'll discuss the theoretical value of copyleft in a world where FOSS has "won", indicated demand for copyleft in the culture and data space (e.g., institutional partners pushing for data in CC 4 & CDLA; interest in non-licensing solutions), License Zero, the legal challenges of copyleft in the SaaS space, and the growing concerns about developer sustainability. [Tidelift, my company, is working in this sustainability space, but obviously I'll avoid making the talk a pitch for the company.]

I will likely conclude (contingent on some further data analysis) that there continues to be interest and demand for copyleft, but that a major driver of the perceived (and in some cases very real) decline of GPL + friends is a result of the inadequacy of our current copyleft legal tools. My hope is that this will be a call to action to the community to continue innovating around copyleft.


Photo of Luis Villa Luis Villa