> "Declarative" always exists in a spectrum: one's specification is another's implementation. I'll go further: statelessness and applicativity obviously have a

> "Declarative" always exists in a spectrum: one's specification is another's implementation. I'll go further: statelessness and applicativity obviously have an intimate and important relationship with declarativity ("what rather than how") (just as non-Turing-equivalence does), but they're not the same, and trying to simply define the latter as the former is an annoying PR wheeze that spreads obfuscation.