5 responses to “Prototype vs. Simulation”

  1. Charles J Gervasi

    I’m reminded of Bob Pease’s suggestion that building something is the best simulator.

  2. Dave

    All simulations are wrong. Some simulations are useful.

  3. Melesio

    Interesting. I never thought about this problem mathematically. It does make more sense and it is a lot more practical to make prototypes instead of trying to obtain an optimal solution straight up from a gigantic amount of simulation possibilities. Great blog post.

  4. ChrisW

    I am not sure I understand the problem being reviewed here.

    The observation that exhaustive testing is expensive, is correct, but if the problem is to reduce the time needed to ‘debug’ prototypes, then I am not sure exhaustive testing is the obvious follow-on.

    To me debugging is the process of removing errors from a solution when the only tool we have available is try-it-and-see-what-happens (beer not included).

    But, if we are working in well a defined problem space, and tools are available to allow 100% simulation, then I would think it better to ‘debug’ a simulation as opposed to building a prototype and debugging that.

    Would I then run an exhaustive test on the simulation? No. I would run design proving tests on the simulation, and compare the results with as built testing on the product.

    My caution would be about the effort needed to derive the simulation. If it is substantively more than producing a prototype (including test harness), then I would stick to the prototype.