> Programs that adhere to these tenets should in theory be far more optimizable than what would be possible in a procedural language, as the compiler is essentially given the entire evaluation graph for the program up front and can completely reschedule evaluations in whatever order is most efficient. They should also be more amenable to formal TLA-type proofs since their computation model is much closer to what TLA+ expects.
This sounds a lot like graph computation but on an extremely granular level.
This sounds a lot like graph computation but on an extremely granular level.