1. 15

  2. 4

    I find thinking of it as “time travel” is more confusing; thinking of it as lazily evaluating just enough to compute the thing we need to compute, a la the Y combinator, made a lot more sense to me.

    1. 1

      The “time travel” model does seem like a pleasingly easy mental model for whether or not the evaluation is going to diverge though.

      1. 1

        Disagree; “time travel” should just work, whereas it’s obvious that lazy evaluation can diverge in some cases, and you can think through whether it will or not.