That's one likely reason Apollo succeeded as well as they did, because they didn't have fancy computers back then.

I gotta register a minor quibble about your characterization of the relationship between computing power and engineering success.

Apollo was developed at the height of the Mainframeozoic Era and the Shuttle in the sunset of that epoch, but the computers used in both projects weren't that dramatically different. Certainly not compared to the mind-boggling explosion of power when the mainframe dinos gave way to the nimble warm-blooded PCs of the Microprocessarian Era. I'll whip out my trusty cliche and flog it yet again: The computer on which I'm pounding out these weighty thoughts is more powerful than the mainframes used to design the Shuttle.

To put the relationship between computing power and engineering in more realistic perspective, plot three points on the timeline--add Burt Rutan and the other X-Prize-generation contenders. You have maybe five years between the peak of Apollo and the beginning of Shuttle development, aybe ten depending on how you define points; and another twenty five years separates the Shuttle from SpaceShip One, spanning the Microprocessarian Explosion boundary. Safe to assume that Rutan and Co. started with CAD and simulation to get quickly into the right ballpark, before they went realworld with wind tunnel and drop tests. Look at what they accomplished for so relatively little money! Simulation isn't the magic bullet, but I do expect it to be powerful force-multiplier when it comes to things like the CEV.

Yet, I don't at all disagree with the anti-simulation sentiments of the thread. I'd trust the protective power of a simulation so far as weeding out the gross problems and helping to get quickly into the right ballpark. But simuilations have little "generative" value in producing a superior result. For that you need human creativity and realworld testing.

Now that I think about it, there's an inherent conceptual flaw in relying on simulation: An accurate simulation is yesterday's simulation. We use simulations to push the envelope, so each new sim inherently embodies uncertain and flawed knowledge, and assumptions about the new territority. Sims will always, by definition, be misleading.

Better simulations will give better results, but at the same time we're aiming higher, so it's almost a wash; it doesn't tip the balance away from the need for realworld testing. Simulation's good, on balance, taken with a grain of salt and lots of skepticisim, with one hand on your wallet...yaaaay simulation...sorta, kinda.