One could also say that functional programming prefers expressions to statements —and treats functions as just another expression. When needed, statement-like t

One could also say that functional programming prefers expressions to statements —and treats functions as just another expression. When needed, statement-like things are constructed from expressions. Equations in lambda calculus do seem to have a "preferred direction" specified by the evaluation rules, though referential transparency lets you go "in the other direction" in your head, when reasoning about your program. Perhaps the programs ~ equations idea would apply even better to logic programming, or term rewriting?