Some years ago by my reckoning, we passed peak GNU. It went by quicker than some of us noticed, but it was anticipated. By peak GNU, I refer to the impact of the GNU project on free and open source software, both in terms of technology and in terms of license choices.
The GNU project, more than any, have pioneered the idea of copyleft licenses through the GNU GPL. It has also been a strong advocate for quality. When people were hired to work for the GNU project in the mid 1980s, the code was scrutinized in excruciating detail for conformance to the coding standards.
When I worked as a webmaster for the GNU project in the end of the 1990s, I experienced the same attention to detail when it came to writing. In my interviews with former staff for the GNU project, one developer noted how he had never, ever, neither before nor after, in all his working career in the industry, seen the same obsession with quality as that which he saw when he joined the GNU project.
But no program lasts forever. Several of the packages which I would consider make up the core of the GNU project increasingly find themselves sidelined: Emacs doesn't have the same appeal as it once did, bash is losing its status as a default shell, gcc has increased competition and the entire tool chain looks different in many projects. What was once GNU, isn't new any more.
While the evaluation guidelines of the GNU project talk about the need for programs adopted by the GNU project to work together, and how GNU is not just a collection of useful programs, I believe the reality is different. Looking at the collection of GNU programs, there are a part of them which are clearly developer focused and serve the GNU tool chain (automake, gcc, etc).
Another part are general programs which you would assume to be part of a normal operating distribution, which the GNU project has always sought to build: dia, gnome, gpaint, and so on.
Then there are the outliers: gneuralnetwork, melting, health, and so on. All worthy programs in their own right, but neither developer tools nor something you expect as part of a standard operating system distribution.
The GNU project has a lot of potential, and a lot of good will. There's a need for projects driving copyleft adoption and showing how copyleft licensing can be a positive force for free and open source software. But there seems to be some ambivalence in how the GNU project presents itself, what programs it accepts as part of the GNU family, and how it is organised.
If nothing else, perhaps a clearer differentiation between the GNU tool chain, the GNU operating system, and additional GNU tools, would be in order.
What would be your thoughts?