Author: Dennis Ritchie (C, Unix)

Dennis Ritchie created C and co-created Unix. His retrospective writing is unusually candid about what C got wrong, which makes him the best primary source on the cost of deferred safety decisions.

Ethos in his own words

“C is quirky, flawed, and an enormous success.” (The Development of the C Language)

“UNIX is a simple coherent system that pushes a few good ideas and models to the limit.” (Reflections on Software Research, Turing lecture 1983)

“Perhaps the greatest danger to the ideal I have described, of research and development groups intimately connected with the technology they are inventing, may be excessive relevance.” (Turing lecture 1983)

“Other issues, particularly type safety and interface checking, did not seem as important then as they became later.” (The Development of the C Language)

“Chief among these is that the language and its generally-expected environment provide little help for writing very large systems.” (The Development of the C Language)

Defining decisions

Types describe machine representation, not intent. C’s type system existed to tell the compiler how to lay out memory and how to generate arithmetic; safety was secondary, by his own admission: “Other issues, particularly type safety and interface checking, did not seem as important then as they became later.” (chist) How it aged: the entire memory-safety industry.

Operator precedence he would not repeat. Ritchie records the & and == precedence error and its cause — the introduction of && after the precedence table had been set, with existing code already written — as a mistake preserved for compatibility (chist). Judgment: the lesson for Mo is that a syntactic decision becomes unfixable the moment a corpus exists, and an agent-written corpus grows faster than a human one.

Small composable tools instead of large programs. The Unix model pushed “a few good ideas and models to the limit” (Turing lecture), which is the ancestor of Mo’s small-surface stdlib and of the shape laws’ intent.

Warning against excessive relevance. His Turing lecture’s named danger is research bent to immediate application (Turing lecture). Judgment: Mo’s risk is the inverse — a design wiki running well ahead of the implementation.

Where he contradicts Mo

Ritchie’s diagnosis is that C lacked help “for writing very large systems” (chist), and his remedy line is module and interface machinery, not line-count and depth limits. Mo’s answer to large systems is currently a set of numeric caps, which bound files rather than structure dependencies.

What Mo could take

ideamaps tostatus
Freeze syntax before the corpus exists, because agents build corpora fastd20-human-pulled-in-when-shape-changesstrengthens Mo
Large-system help means interfaces and dependency structure, not only size capsq12-law-numberscontradicts Mo
Push a few models to the limit rather than adding mechanismsd27-simple-and-elegant-like-rubyalready in Mo
Publish a “what we knowingly deferred” list, as chist doesd19-negative-space-is-the-contractnew idea for Mo

17 Sep 2026: the numeric-caps row was ruled agree, already in Mo on 13 Sep — see research-agenda-2026-09-response.