Chapter 12: The Principle of Computational Equivalence

Section 8: Undecidability and Intractability

Halting problems

A classic example of a problem that is known in general to be undecidable is whether a given Turing machine will ever halt when started from a given initial condition. Halting is usually defined by the head of the Turing machine reaching a special halt state. But other criteria can equally well be used—say the head reaching a particular position (see page 759), or a certain pattern of colors being formed on the tape. And in a system like a cellular automaton a halting problem can be set up by asking whether a cell at a particular position ever turns a particular color, or whether, more globally, the complete state of the system ever reaches a fixed point and no longer changes.

In practical computing, one usually thinks of computational programs as being set up much like the register machines of page 896 and halting when they have finished executing their instructions. User interface and operating system programs are not normally intended to halt in an explicit sense, although without external input they often reach states that do not change. Mathematica works by taking its input and repeatedly applying transformation rules—a process which normally reaches a fixed point that is returned as the answer, but with definitions like x = x + 1 (x having no value) formally does not.

Image Source Notebooks:

From Stephen Wolfram: A New Kind of Science [citation]