Qualitative behavior: KISS me

HydroComp Web Log

[efficiency in motion for over 20 years]

Qualitative behavior: KISS me
Miscellaneous Musings from the Technical Director

Monday, 11-Feb-2008 by Donald MacPherson - HydroComp Technical Director

Which would you rather have - a weather report that was perfectly correct 80% of the time and completely wrong the remaining 20%, or one that was exactly opposite 100% of the time. How about a hydrodynamic code that was within 5% for most of your hulls, but had the potential for over 50% error on some portion of your designs? Would you rather have a method that was universally within 15% for all of your hulls? Give me the weather report that is always wrong, and I'll plan for the opposite. Give me the code that was always within 15% and I'll develop my own correlations. The first objective of any engineering analysis must be to avoid the big errors. In my opinion, consistently good qualitative behavior always trumps inconsistent quantitative accuracy. So how does one get there? KISS me.

You remember - KISS..."Keep it Simple, Stupid." Far too often, people look to increasing complexity for "improvements" in design. Consider how complex cars have become. My daughter has had two iPods that have developed software problems and die - is a device like an iPod really that much better than a simple portable CD player? I cringe when I see engineering students equate added complexity with increased "client benefit."

And, of course, just look at how hydrodynamic codes have evolved, from basic potential flow, to RANS methods, to BEM codes, to who knows what comes next. Each generation is made more complex to solve a narrower problem. So long as the scope of the problem remains small, the codes do indeed provide better solutions - but at what cost? They can only be used by experts, and even then, there is a bit of apprehension in the validity of the results. Why can't we have really basic hydrodynamic codes, that are fundamentally simple, universally well behaved, and useable by the practicing naval architect? (Maybe its time for HydroComp to do something about this? We'll see...)

And, I always wonder if weather reporters ever actually look out the window before making their broadcast...