[ClearCase_guy]

I won't say that all computer simulations are worthless, but I think most of them are.

Let's say I'm writing a SW simulation of a tea kettle on a stove. I finish and run the simulation -- but my water never boils. Obviously I screwed up. So I re-write my simulation. I keep running, testing, and re-writing my simulation until I see that (according to my model) water in a tea kettle will boil when placed on a hot stove.

What has my simulation showed me? Absolutely nothing. I made sure that it showed me what I wanted it to show me. That's how I knew it was "successful".

Global Warming? Asteroid strikes? If the initial run shows a really happy outcome, then your simulation was badly written. So keep re-writing it until Global Warming or Asteroid Strikes show a catastrophic outcome. That's how you know your simulation is "good".

This is how we do science these days.

[TexasGunLover]

I won't say that all computer simulations are worthless, but I think most of them are.

We find that almost all are perfect representations of real work physics. We run thousands a year, and run dozens of physical tests to confirm their results.

[pfflier]

That is an awesome summary. I have been arguing the same points you made with the algore acolytes for years.

They don't seem to understand assumptions, multiple independent variables and expected outcome, just dogma.

[Yossarian]

Simulations are a key part of designing and verifying electronics. In that realm, they work very well.