What Naur got wrong from The concept of mind

In Peter Naur's essay "Programming as Theory Building," he attempts to construct a definition of programming based on Ryle's concept of theory. However, Naur mistakenly interprets Ryle's notion, leading to a convoluted paper. One of the initial issues is Naur's assertion that the concept of theory, as defined by Ryle, is an "unembodied World 3 object" and therefore has a valid philosophical standing.

While it is true that theorizing is a cultural activity, Popper's introduction of three worlds - material, mental, and cultural - demonstrates that the mind does not exist separate from the physical. Ryle's work, "The Concept of Mind," aims to refute the existence of a mind independent of the brain. Instead, Ryle seeks to clarify the logical structure of existing knowledge.

Naur fails to grasp this distinction, conflating intelligence and intellect and conflating theory with intellectual operations. Ryle, on the other hand, distinguishes intellectual powers and performances from the intellectual process itself, emphasizing that theory-building is one activity among many. Naur incorrectly claims that theories cannot be reduced to rules and criteria, whereas Ryle argues that intelligence is not solely about satisfying criteria but about applying them, detecting and correcting errors, and improving upon successes.

The author contends that Naur's perspective is shortsighted because programming is indeed theory building, even if the programmer is paid for the program rather than the theory. Historians write accounts of battles, which are their theories, and the same applies to programming. However, Naur's primary concern with theory building stems from the need for program modification.

Many software applications have already been written and rewritten numerous times, so starting from scratch is unnecessary in many cases. There are, of course, new and improved theories that may arise, but practical software building relies on established theories. The final point of confusion for Naur involves the process of acquiring a theory, specifically the claim that "program revival" – re-establishing a program's theory from documentation alone – is impossible.

This assertion fails to consider the value of collaboration in learning about a new system or program.