# John Backus

> 1924–2007 · Computer Scientist, Creator of FORTRAN
>
> **Recorded contribution:** FORTRAN; BNF; FP (functional programming language); Turing Award lecture on freeing from von Neumann style

## How to use this dossier

Read for a causal chain, not a hero story: inherited problem → contribution → mechanism → downstream capability → limit. Then close the page and complete the reconstruction exercise from memory.

## 1. Historical orientation

John Backus (1924–2007) led the IBM team that created FORTRAN, one of the first high-level languages to achieve broad practical adoption because its compiler produced code competitive with hand-written assembly. He also helped define the notation later called Backus–Naur Form for ALGOL syntax and proposed the functional language FP in his 1977 Turing lecture. These are related attempts to raise programming above machine detail: compile mathematical formulas, define language structure precisely, and compose programs from higher-level functions. FORTRAN was emphatically a team achievement; Backus's historical role was technical leadership and a persistent challenge to the von Neumann style.

## 2. The problem inherited

Scientific programmers spent enormous effort translating formulas into machine instructions and distrusted high-level languages because inefficient generated code would waste scarce computer time.

## 3. The central contribution

Backus led a compiler project that made formula-oriented source practical through aggressive translation and optimization, then advanced formal syntax notation and functional composition as further abstractions.

## 4. Reconstruct the mechanism

1. Parse algebra-like statements, loops, branches, arrays, and subprograms into an internal representation.
2. Analyze expressions and control flow to select instructions, reuse registers, and reduce loop overhead.
3. Resolve symbolic variables and storage layout, then emit and link target-machine code for the IBM 704.
4. Validate performance against hand coding so reduced programmer effort does not impose an unacceptable execution penalty.

## 5. What changed downstream

- FORTRAN made high-level compiled programming credible and became a durable language for scientific computing.
- BNF-style grammars became a standard way to specify programming-language syntax.
- Backus's functional-programming critique influenced research on algebraic program composition and avoiding stateful bottlenecks.

## 6. Attribution, limits, and uncertainty

- The FORTRAN team included Harlan Herrick, Peter Sheridan, Roy Nutt, Irving Ziller, Lois Haibt, and others whose work should be named.
- Early FORTRAN traded portability and modern safety for performance on a specific machine.
- BNF describes syntactic form, not full meaning, resource behavior, numerical accuracy, or correctness.

## 7. Reconstruction lab

Compile a small array-sum loop by hand into a toy instruction set. Add induction-variable and common-subexpression optimizations, count instructions and memory accesses, and compare with a deliberately naïve translation. Explain what evidence would have persuaded a 1950s assembly programmer.

## 8. Evidence trail

- [The History of FORTRAN I, II, and III](https://doi.org/10.1145/800025.1198345) — ACM SIGPLAN
- [Can Programming Be Liberated from the von Neumann Style?](https://doi.org/10.1145/359576.359579) — Communications of the ACM

---

*Research checked 2026-08-09. Dates, roles, and claims about living people are historical snapshots. Linked sources remain the authority; this dossier is original instructional synthesis.*