The Heisenberg Uncertainty Principle places impassable barriers to classical empiricisim. However I suspect there is a fair amount a deep structure to the universe that is unmeasurable. The only way models of this deep structure could be evaluated is through simulation and then observing consequences of the simulation that rise above the limits imposed by Heisenberg's Principle. To some extent this approach is heresy to pure science which states that if a phenomena isn't susceptable to repeatable direct observation then it isn't real. Also there is the issue of multiple solution paths, e.g. many different simulations based upon different assumptions could yield the same observable result.

The problem of different solution paths has been a big problem with heat shield design. The Galileo Probe during its entry into Jupiter's atmosphere produced an expected result. The instruments on the heatshield only measured heat shield ablation as a function of time. Ablation in the Probe's fustrum was much higher than expected (it almost failed). There are two wildly different theories that can produce the same ablation rates (both theories have unknown parameters that can be tweeked). If we incorrectly assume the wrong theory is correct and go back to Jupiter with a heat shield based upon that theory, the vehicle could burn up.

Gary