2 Replies to “Managed Copy & Paste”

  1. There’s no question that copy/paste is widespread. I often google for
    things, copy them from stack overflow and then slightly modify them.

    For the first 2 minutes of this video, I was bothered by the fact that
    the potential values of the stoplight were not well-typed: strings
    were used for something where an enumerated type was needed. As I
    write this, I begin to think: “In Haskell, change-light-fast would
    have a subtype consisting of only RED and GREEN to further tighten up
    the code” … in other words, how robust is this code? Does it need
    unit tests?

    The next thing that became obvious that some form of graphical
    programming indicating state transition is far more articulate than
    text code in this case. Nonetheless, the need to synchronize delta
    between the visual versions exists also.

    Finally, when you started to create the general abstract version of
    change-light that could be concretized into specific functions, I
    began to think: “He certainly needs code as data. Perhaps using Lisp/Prolog
    is the best way forward instead of this Pascal-ish language he is
    using.” Prolog is especially appealing because the function signature
    makes all potential state change explicit right in the header of the
    function.

    Interesting idea nonetheless.

  2. Provocative as always. How often we face the tradeoff between maintaining near duplicate code or abstracted code complicated by conditionals. What if tools supported both projected partial evaluations as well as the complete parameterized general case?

Comments are closed.