So if someone makes a physical invention that does something they can patent it but if it's something a general purpose computer can be programmed to do, writing that program does not cause the computer to violate the patent.
> My suggestion is to change the effect of patents. We should legislate that developing, distributing, or running a program on generally used computing hardware does not constitute patent infringement.
I guess the phrase "Software Patents" and the strange rules about software patentability in Europe gave me the impression that they wanted to take the non-patentable route instead of the non-infringing one.
In short, I would answer your question with a question. Why can hardware violate a patent and software not?
Going from your example, in the absence of cheap programmable computers the problem of coordinating spark timing with the engine is a real engineering problem that probably required nontrivial R&D resources to get right. To the extent that distributors remained useful to the world, those sunk costs should be protected. If you didn't, there'd be a toxic second-mover advantage to just copying the other guy's R&D work.
Software just doesn't have those kind of sunk-cost discovery problem, at least not at anything like the scale other disciplines do. The fixed costs in software are different. They're almost all about execution, education, coordination, and marketing.
Imagine a software R&D guy, who discovers a new thing you can do with software. Assume the idea works, and is profitable: He's still spent a tiny fraction of bringing a product to market, even a product that is just an implementation of his invention. Patents won't fix the second-mover advantage, if any, because the first-mover costs aren't from software research.
The mismatch actually starts to matter when non-practicing inventors come into play. In the old world of industry, these guys were at least providing a public service by creating inventions nobody else bothered to fund, so it was reasonable that companies that found their inventions useful should pay for them. But in software the non-practicing inventor hasn't paid the cost. There is still the first-mover disadvantage out there of building the market, educating your users, etc. on top of the cost of turning the invention into a shipable product that everyone in the market has to pay.