pseudocode

Oliver Byrne (1847) as a Pointer-Free Visual Programming Language

Long before structured programming, dataflow visual languages, or modern IDE token highlighting, Oliver Byrne (1847, pp. ix–xi, xxvii–xxix) diagnosed the fundamental cognitive bottleneck of classical Euclidean deduction: alphabetic pointer dereferencing.

Pass-by-Reference vs. Pass-by-Value in Geometry

In Euclid's Greek Stoicheia and Heath's English edition (geometor/euclid), geometric primitives on the page are nameless black ink strokes. To refer to a segment, angle, or square, prose assigns memory addresses ($A, B, C, D, \dots$) to point vertices and forms composite tuple expressions (AB, angle FBC, parallelogram BLMD). Reading a single sentence of I.5 or I.47 requires dozens of indirect table lookups:

\[\text{String Token } (\text{``}\triangle FBC\text{''}) \;\xrightarrow{\text{scan}}\; \{F, B, C\} \;\xrightarrow{\text{trace}}\; \text{mental hull} \;\xrightarrow{\text{compare}}\; \text{String Token } (\text{``}\triangle ABD\text{''}).\]

Byrne’s 1847 Chiswick Press grammar replaces pass-by-reference with pass-by-value:

  1. Typed Inline Value-Tokens: Every primitive sort—segment (), angle (), triangle (), square (), rectangle (), arc (), circle (), and Book V magnitude (, , , )—appears directly inside the algebraic line as a miniature colored geometric literal.

  2. Formal Operators (pp. xxvii–xxviii): Natural-language verbs collapse into symbolic operators: = (equality of length, angle, or area), + / - (composition and subtraction), ⊏ (greater than), ⊐ (less than), ∥ (parallel), ⊥ (perpendicular), (one right angle), (two right angles), ∷ (proportion), ∵ (because), and ∴ (therefore).

  3. Structured Proclean Control Flow: Every proposition is formatted into explicit visual blocks matching Proclus's six divisions (ekthesis, diorismos, kataskeue, apodeixis, symperasma), with each deductive equation placed on its own line accompanied by marginal :just: citations.

Quantitative Rosetta Compiler Metrics

The module geometor.byrne.pseudocode parses seven flagship propositions across Books I–V and measures verbal token load, alphabetic pointer dereferences, and visual register pressure:

Heath (1908) vs. Byrne (1847) vs. GEOMETOR DAG Metrics

Prop

G-Index

Heath Words

Heath Pointers

Byrne Words

Verbal Load

Visual Registers

I.1

G.84

179

31

22

12.3%

6 (unique)

I.5

G.91

329

62

32

9.7%

8 (+4 congruence twins)

I.47

G.140

314

64

25

8.0%

9 (stroke-modulated)

II.11

G.154

269

56

50

18.6%

5 (solid-dotted cut)

IV.10

G.223

389

78

38

9.8%

9 (composite + dotted)

IV.11

G.224

210

41

42

20.0%

10 (5-side register overflow)

V.16

G.276

253

20

18

7.1%

4 (pure Eudoxian algebra)

Three Structural Laws Discovered by Register Analysis

  1. The Four-Color Working-Memory Ceiling (:term:`Register Pressure <register-pressure>`): Human visual working memory holds $4 \pm 1$ chromatic chunks effortlessly. Whenever a proposition introduces five or more primitives of the same sort—such as the five sides of the regular pentagon in IV.11 (, , , , ) or shared square/rectangle boundaries in I.47—Byrne multiplies his four ink channels by stroke modulation (solid, dashed, dotted, solid-dotted).

  2. Intentional :term:`Congruence-Twin Coloring <congruence-twin>`: In I.5 (Pons Asinorum), Byrne deliberately binds both equal legs to , both produced extensions to , and both auxiliary cross-diagonals to . What a naive type-checker would flag as a register collision is in fact a symmetric congruence invariant encoded directly into the visual type!

  3. Why GEOMETOR Pairs Visual Tokens with Symbolic DAGs: Pure visual tokens excel at human proof comprehension (apodeixis), whereas machine constructibility requires explicit incidence/orientation selectors (side_of, beyond) when circles intersect in two points. Together, geometor/euclid (textual heritage), geometor/byrne (pointer-free visual cognition), and geometor/elements + geometor/codes (computable symbolic DAGs) form a complete three-pillar Rosetta Stone.

See Unit 7 of the Book Codex for full side-by-side 3-way derivations of each proposition.