> I've spent a lot of time perusing various open-source code bases, specifically OS kernels (FreeBSD and Linux, mostly), and I have been pretty dismayed to find

> I've spent a lot of time perusing various open-source code bases, specifically OS kernels (FreeBSD and Linux, mostly), and I have been pretty dismayed to find far too much raw egotism/intentional obscurity, frankly lazy hacks, and poor documentation. Delving into user-land libraries can be down-right terrifying. It's not a problem of ineptitude, it's a combination of over-confidence, and the weakness of mainstream languages to properly abstract systems, and contain side-effects. When I was younger, I use to think I just wasn't "advanced enough" to understand what I was looking at. After becoming experienced, what I really found was that poor practices, both in design and implementation, are endemic in mainstream software. > The result was 99% Ada (Ada 2012-FSF GNAT-FreeBSD). I mean 99% as in I didn't use any external libraries. The only non-Ada components were some last-mile system-calls bindings written in C, to take advantage of the system headers. All JSON parsing/generation, HTTP, and TCP/IP was implemented in Ada. I think implementing your own software from end to end (including things like HTTP and TCP/IP) in Ada (a relatively obscure language) is "far too much raw egotism/intentional obscurity". The author seems to think he has solved the problem of collaboration by working alone on some software on its own terms.