In reply to jimtitt:
> Such are the difficulties in life and why a theoretical approach has realistically been abandoned, the approach some of the CAI guys have taken is the better one which is to find a model which fits the observed data.
May I split some hairs? (Sorry, an occupational proclivity.) All the approaches to modeling are fundamentally theoretical. The equations are not the rope, they are a mathematical construction. Moreover, every model from the simplest, the poor much-maligned standard model, through an increasing range of less inaccurate ones (I'm choosing double negatives intentionally), require some input from observed data. The CAI guys have strung together a bunch of modules, one for each element of the belay chain, each one of which is the good old "theoretical" standard model with an equally "theoretical" viscous damping term added on. These things are classical theoretical damped harmonic oscillator terms. Then they've added terms to describe friction losses between each point in the belay chain, and yes, having created a host of independent parameters, now require a large amount of experimental data in order to determine those parameter values. And yes, now that their model is a big system of differential equations, they cannot provide an analytic solution, instead requiring a computer to implement standard numerical analysis techniques in order to get answers, which apparently agree very well with real-world results.
Did they "realistically abandon" a "theoretical approach" to do this, or rather, did they implement a considerably more complex theoretical approach capable of taking into account more features of the real-world situation? I'd vote for the latter view.
This is, of course, a matter of philosophy rather than any kind of practical distinction. The CAI guys got a better fit, and they needed a big load of data to get it---but they got a better fit. But I can't detect any "theory" abandoned by the roadside.
I mention this, if there is anyone still listening and still interested, because it can happen that a data-fitting approach gives us predictions but no understanding, where by "understanding" I mean the sense that the phenomena we observe can somehow be seen to be consequences of some set of fundamental truths. This is Newton's paradise, and although we may have neither the time nor the ability to dwell there, some of us anyway long for a path back, whereas others want useful answers now that they can employ in the design of better stuff.
Ok, back to your regularly schedule programming...