Simulate to iterate

Run the work as a cheap, repeatable experiment.

Search the space of setups by simulating the work, not by doing it for real. Iteration is only leverage if it is faster and cheaper than the work itself.

The test: Can you rehearse this work faster and cheaper than doing it for real?

Why

A hard problem has thousands of configurations: which model, which tools, which representation. Trying each in production costs a person-week per point, so nobody explores; they commit to one setup and defend it. Simulation collapses the cost per point, which turns design from argument into search. Without it, "iterate" is just guessing slowly.

How we apply it

• sysml-bench rehearses structured-engineering work as N-trial experiments, so a week of an engineer's task runs in an afternoon of trials.
• The configuration space (model x tools x representation) is traversed in simulation; only the survivors reach production.