3 ms·
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 f
by dschiptsov 10y ago
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.)
- TeMPOraL 10y agoAll 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.
- incepted 10y ago> 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.
- sjy 10y agoGood polysemous use of the word 'mutations.'