Sussman cites a desire to mimic biology in his essay "Building Robust Systems"

https://groups.csail.mit.edu/mac/users/gjs/6.945/readings/ro...

It is actually a long story, and it goes back to the old dogma that lists and structures made out of conses are good-enough, almost universal building blocks for code and data. Type-tagging was obviously inspired by DNA, etc.

As for protocols, there is, for some obvious reasons, still nothing better than asynchronous, out of order message-passing over a packet-based networks, with fixed-size header (upon which one could patter-match in a single pass) and adjustable size payloads, which could adapt to a physical channel by adjusting to its frame-size, so, the stdlib and OTP is good-enough and well-balanced and just works.

It is also not a coincidence that Erlang is pure-functional - so is the world of proteins and enzimes.

There is a bit more universality in a chunked linear structures and stateless asynchronous message passing of type-tagged binary data with feed-back loops co-regulation that it seems.)

All things you mention are library-level stuff. It's also no surprise that Erlang is actually a Lisp with Prologish syntax - just take a look at the intermediary language that sits between Erlang and BEAM bytecode ;). Yeah, the one you have to learn about if you want to do parse-transforms.

> Type-tagging was obviously inspired by DNA

Source?

> It is also not a coincidence that Erlang is pure-functional - so is the world of proteins and enzimes.

That world is mostly driven by mutations, which are the exact opposite of functional purity.

I really have no idea where you get all these broken metaphors.

Good polysemous use of the word 'mutations.'