Over the last 2-3 years, we have been noticing an interesting trend in the software world. GPL is dead and MIT/BSD/Apache-2 types of licenses are spreading everywhere. Only when we come to the academic world of bioinformatics, we find large majority of codes still being released under one or other form of GPL.
We have been observing the death of GPL everywhere - at github, in various software communities and in languages and libraries newly released by companies/academia.
1. For example, you can check the most downloaded PHP packages here. None of them is released with GPL any more.
2. Or check the license for code of Google’s GO language.
3. If you argue that Google is too big and can afford to throw a few freebies, check the license of Julia language released by three programmers.
4. We have been checking all over github for highly used programs and packages, and we notice the same pattern - GPL is dead and MIT/BSD are everywhere.
What is going on? To find out, we asked Titus Brown, one of the rare non-GPL bioinformaticians for his reason to use BSD over GPL.
I have one simple reason: BSD maximizes use. I have lots of background in OSS (going back to 1988 or so) and while I personally agree ethically with GPL, as a publicly funded scientist I believe that I should be maximizing the utility of what I do. This clearly indicates a non-contaminatory license.
Titus is correct. GPL is avoided by developers, because it restricts sharing and collaborative projects and imposes the will of the GPL-licensor on the others. Phil Sturgeon, a highly skilled web-developer, shared the following figure in his twitter channel describing the evolution of web technology -
Such extensive sharing of libraries and code would not have been possible without each library being released under BSD/MIT-type license.