Home > Uncategorised > Here’s A Question

Here’s A Question

January 17th, 2006 Leave a comment Go to comments

Are all software bugs repeatable? As in, if you discover a bug in a piece of software, is there any type of bug that could not be repeated?

My answer is “Yes, all bugs are repeatable”. Bugs depend on circumstances, so as long as you can repeat the circumstances, then you can repeat the bug.

I’m betting Richard thinks otherwise…. 😉

Categories: Uncategorised Tags: , ,
  1. Rich
    January 18th, 2006 at 09:48 | #1

    In a way
    Consider:

    – Introspective software, where the bug depends on prior bugs, sequencing, or the introspection itself. If you object to this one, consider one of the previous bugs itself being temporally sensitive.

    – Range-based bugs revealed by stochastic inputs, where those inputs necessarily arise from an external source (such as an atomic source).

    – Any program whose state reflects or includes the environment, and that environment is mutable or consumable. Good luck testing your printer driver bug with no paper (or a software condition that only arises from incorrect handling of a hardware fault). Obviously, printers are a bit of an exaggerated case, but I’m sure you get my point. Simulated interfaces are all well and good, but might be insufficient.

    In conclusion — I mostly agree with you. GP computers can emulate any other GP computer, so pure software can always have its execution environment perfectly replicated — *in theory*. Where practice is necessarily different, or the computer is not replicable (as in, say, some kind of "world computer"), then I disagree.

    Of course 🙂

  1. No trackbacks yet.