> the number of useful libraries
So how do you decide what's useful? I just wrote a Python library that talks the Bluetooth protocol from my Zero electric motorcycle. I mainly use it to monitor how charged it is from the other end of the house.
That's surely a library that's only useful to a handful of people on the planet, but it's still useful, and it's on GitHub in the hopes they stumble across it.
They're ALL "useful" because nobody would spend the time writing something without a use, aside from Microsoft and Oracle.