[wizeman]

> You answered your question as you asked it. If you juice up your Turing machine like that, it will still suck so much that nobody will actually write programs for it, but it will be more complicated to describe and analyse.

Of course. Why would you do that? You don't have to.

You can write a program in a Turing-complete language and yet analyze how it will run on a machine with a finite amount of memory.

For example, when you analyze cycle detection algorithms, which assume that the program being analyzed runs on a machine with a finite amount of states (even if it is implemented in a Turing-complete language), you don't do any of that.

And the currently-known algorithms are very simple to analyze.