@cwebber

I knew a fellow who had this conversation with someone in the Rand Corporation, in the 1960s perhaps.

The way he described it, there were two floors, and he worked on the first floor, but one day he went up to the second and asked what they were working on. He explained why it couldn't be done, and then went back down to the first floor.

(Another fellow worked for the Pentagon and got questions to answer in units of GDP. But nobody was reading the answers, fortunately.)

@cwebber My (limited) understanding of the halting problem is that it's only proven to be undecidable because it's defined in a way that makes it so, i.e. by allowing the tested function to run the testing function on itself and doing the opposite of the result.

I'm not saying it's easy to write code to determine whether a function will halt, but if a human can use logic to make that determination, then a machine should be able to make that same determination by running the same algorithm.