A computer is not a box invented by one person on one day. Written procedures read by people, hand-carried digits, state categories punched into cards, logical expressions, opening and closing relays, and instructions in memory acquired new material forms in different cities and institutions.
WHAT YOU SEE ON THIS RIVER
- Diagram in the sky
- The flowchart of a repeating procedure
- Emblem at the source
- Meshing gears
- The real place around each stop
- Around each stele the land takes on the natural geography of that scene’s real place — sea or lake, plain, hills or mountains, the colour of the ground and its common trees — and, where one defines the place, its landform: a volcano, snow peaks, granite domes, a mesa, dunes, a fjord, islands, a rock hill, a gorge or loess terraces. The water near the stop takes the colour of the real river or sea, and the haze the place’s climate. A small globe on the stele marks where it is, with the route from the previous place. Where a city has an iconic building that already stood in the scene’s year, its schematic silhouette rises behind the stop and is named on the card. The land follows today’s terrain and climate as a sketch and the silhouettes are not measured reconstructions. Between stops the river itself stays symbolic.
- A figure board at every stop
- Each board draws the mathematics of that scene. When the boat arrives, the construction is drawn in and the key result rises in red. The drawings are schematic reconstructions, not historical manuscripts.
- Century bands along the banks
- 500–1449 · Stone stele · paper lanterns · single-arch bridge · watermills and villages · lateen boats
- 1450–1749 · Marble stele · iron lanterns · three-arch bridge · windmills and clock-tower towns · sailing ships
- 1750–1899 · Cast-iron plaque · gas lamps · iron truss bridge · factory chimneys, railway and steam train · steamboats
- 1900–1969 · Concrete marker · electric streetlights · concrete bridge · pylons, apartment blocks and radio masts · barges · aircraft
Where the century band changes, the boat passes under a bridge of the new band. Villages, mills, factories, pylons and towers stand for the technology of each century, not for any real place or architectural style.
01·c. 825 CE·Baghdad(basis: Composition)
Al-Khwarizmi — Writing Calculation as a Procedure
Al-Khwarizmi organized arithmetic with Indian numerals so an Arabic reader could follow the operations in order. The original is lost and its exact date unknown, but his name survived in Latin as Algoritmi and became the source of “algorithm.”
- Pause and ask
- Can a skilled calculator's practice be written as an order another reader can reproduce without that person present?
- How thinking changed
- By arranging operations with Indian place-value numerals into ordered prose, arithmetic became not only an answer but a repeatable procedure. This was not a modern programming language or the later abstract definition of an algorithm, but it made rules followed by a person into a transferable unit of knowledge.
- What we cannot claim
- The original Arabic arithmetic is lost and its exact date is unknown, so c. 825 is an editorial anchor. This does not mean al-Khwarizmi single-handedly invented every algorithm or decimal numeral, or that he wrote modern code.
- This place
- The Abbasid court, administrative calculation, and cultures of translation and copying in ninth-century Baghdad created conditions for comparing numeral systems and astronomical and arithmetical texts. The work is not assigned wholesale to a single building called the House of Wisdom. (Tigris banks · date palms · flat plain · 33.3°N 44.4°E)
- Figure board
- Following numbered lines of prose, two numbers in place-value columns are combined right to left, carrying over, to reach the answer.
- On the river
- A desk holding a written record · 500–1449 (Stone stele · paper lanterns · single-arch bridge · watermills and villages · lateen boats)
02·c. 1150 CE·Toledo(basis: Translation)
Latin Algoritmi — A Name Becomes a Method
Across twelfth-century Iberian translation networks, al-Khwarizmi’s arithmetic met new readers through several Latin versions and adaptations. The c. 1150 date is an editorial anchor, not a claim that one Toledo institution, translator, or moment delivered numerals and procedures to all Europe.
- Pause and ask
- How did a person's name become, in another language, a name for a learnable method of calculation?
- How thinking changed
- Latin readers translated, adapted, and annotated unfamiliar numerals and operations, reading *Algoritmi* both as an author's name and a manner of calculation. As procedures became teaching texts in a new language, a proper name gradually moved toward a general word for method.
- What we cannot claim
- The c. 1150 date is an anchor for several textual traditions. It does not claim that one Toledo translator rendered one complete manuscript and delivered numerals and algorithms to all of Europe at once.
- This place
- The multilingual environment of twelfth-century Toledo and Iberia brought Arabic scientific texts, Latin readers, and the language skills of Jewish, Christian, and Muslim scholars into multiple translation relationships. It was not one centrally organized modern translation school. (Tagus bend · river gorge · olives · 39.9°N 4.0°W)
- Figure board
- Several versions branch from a lost original into a network, and following the procedure in any copy gives the same answer.
- On the river
- Two books facing each other · 500–1449 (Stone stele · paper lanterns · single-arch bridge · watermills and villages · lateen boats)
03·1642 CE·Rouen(basis: Main activity)
The Pascaline — Delegating Carrying to Gears
At nineteen, Pascal built a machine in Normandy whose gears performed addition, subtraction, and automatic carrying to help with his father’s tax work. Surviving machines show that part of manual arithmetic could become physical motion; they were not mass-produced general-purpose computers.
- Pause and ask
- What changes when carrying between digits moves from a person's memory into the motion of gears?
- How thinking changed
- A wheel completing one revolution advanced the next position, turning carrying into a physical state transition. With part of the rule fixed in the mechanism, a user handled input and output instead of remembering every intermediate carry.
- What we cannot claim
- 1642 anchors the early construction, not necessarily the date of every surviving machine. The Pascaline automated parts of addition and subtraction but was neither a mass-produced programmable computer nor asserted to be the absolute first calculating machine of every kind.
- This place
- The repeated accounts handled by Pascal's tax-collecting father in Rouen supplied a concrete problem for mechanizing reliable addition and subtraction, while precision metalwork and patent patronage enabled multiple prototypes. (Seine meander · broadleaf trees · valley hills · 49.4°N 1.1°E · landmark: Rouen Cathedral (1506))
- Figure board
- When the units wheel completes a turn from 9 to 0, a catch pushes the tens wheel one step (+1): carrying becomes motion.
- On the river
- A place of ongoing work, marked only by the route’s emblem · 1450–1749 (Marble stele · iron lanterns · three-arch bridge · windmills and clock-tower towns · sailing ships)
04·1834 CE·London(basis: Main activity)
Babbage — Designing a Machine for More Than One Table
After the Difference Engine project stalled, Babbage began conceiving an Analytical Engine around 1834, with punched-card instructions, a store for numbers, a mill for operations, and repetition and branching. The question shifted from one table-making machine to a design accepting many procedures, but the complete engine was never built in his lifetime.
- Pause and ask
- Could a machine accept different procedures on cards instead of being built to produce only one kind of table?
- How thinking changed
- Babbage distinguished a *store* for numbers, a *mill* for operations, punched-card instructions, repetition, and conditional branching, separating machine architecture from the procedure performed. Different instruction arrangements, rather than permanently fixed gearing, could express different calculations.
- What we cannot claim
- 1834 marks the emerging Analytical Engine concept, not one completed design date. Because the full machine was neither built nor run in Babbage's lifetime, architectural resemblance to a modern computer is kept distinct from demonstrated operation.
- This place
- London's Royal Society, government funding, precision engineers, and demand for printed tables made a large calculating-engine project possible, while cost and conflict among Babbage, government, and makers blocked completion. The same city exposes conditions of both design and failure. (Thames banks · broadleaf trees · flat basin · 51.5°N 0.1°W · landmark: St Paul’s Cathedral (1710), Tower of London (1100))
- Figure board
- A chain of punched cards directs columns that hold numbers and a wheel that operates, with a loop back (repetition) and a fork (branching).
- On the river
- A place of ongoing work, marked only by the route’s emblem · 1750–1899 (Cast-iron plaque · gas lamps · iron truss bridge · factory chimneys, railway and steam train · steamboats)
05·1843 CE·London(basis: Publication)
Ada Lovelace — Seeing Machine and Program as Distinct
Lovelace translated Menabrea’s account of the Analytical Engine and added notes more than three times as long. Note G’s Bernoulli-number table is often read as the first published computer program, though collaboration with Babbage and definitions of “program” remain debated. Her clearer leap was to distinguish rules that might manipulate symbols, letters, or music as well as numbers.
- Pause and ask
- Can what an unfinished machine should do be described as a table of instructions distinct from its parts?
- How thinking changed
- Lovelace's Note G arranged operations and intermediate variables for Bernoulli numbers in a table, inviting readers to distinguish the machine from its procedure. More broadly, she argued that relations expressed symbolically might let rules operate on letters or music as well as numbers.
- What we cannot claim
- Note G is widely read as the first published computer program, but definitions of program, Babbage's earlier tables, and the division of collaboration remain debated. Lovelace is neither made a lone inventor nor erased as a mere copyist.
- This place
- London scientific networks, sustained collaboration with Babbage, translation of a French paper, and publication in Taylor's *Scientific Memoirs* turned an unbuilt machine into a publicly discussable text. Class privilege enabled education while formal scientific careers remained narrow for women. (Thames banks · broadleaf trees · flat basin · 51.5°N 0.1°W · landmark: St Paul’s Cathedral (1710), Tower of London (1100))
- Figure board
- A table of operations and the variables they use stands apart from the unbuilt machine; its rules might reach numbers, letters and music.
- On the river
- A stack of books · 1750–1899 (Cast-iron plaque · gas lamps · iron truss bridge · factory chimneys, railway and steam train · steamboats)
06·1854 CE·Cork(basis: Composition)
Boole — Calculating with Forms of Reasoning
As professor in Cork, George Boole treated combinations of propositions and classes through symbolic operations in The Laws of Thought. The book did not design an electronic computer; it made forms of inference objects of calculation. The bridge from logical expressions to open and closed circuits appeared eighty-three years later in Shannon’s work.
- Pause and ask
- Can forms of reasoning such as AND, OR, and NOT be combined and calculated like quantities?
- How thinking changed
- Boole treated combinations of propositions and classes through symbols and laws of operation, separating logical form from content. A logical expression was not yet a wire, but making inference algebraically manipulable and checkable created a language that could later meet circuit design.
- What we cannot claim
- Boole did not design an electronic computer or modern binary logic gates. Eighty-three years of reinterpretation and independent engineering work separate the symbolic logic of 1854 from relay circuits in 1937.
- This place
- A professorship at the newly founded Queen's College Cork gave the largely self-taught Boole time to teach and research, while London and Cambridge scholarly publishing networks connected work written in Cork to wider audiences. (River Lee · broadleaf trees · hills · 51.9°N 8.5°W · landmark: Shandon Steeple, St Anne’s Church (1726))
- Figure board
- The whole, 1, is split into x and 1 − x, y and 1 − y; the part xy is computed, and laws like x² = x make reasoning calculable.
- On the river
- A desk holding a written record · 1750–1899 (Cast-iron plaque · gas lamps · iron truss bridge · factory chimneys, railway and steam train · steamboats)
07·1890 CE·Washington DC(basis: Main activity)
Hollerith — Counting a Population with Punched Cards
The 1880 U.S. census had taken almost a decade to tabulate. In 1890 clerks transferred age, sex, race, citizenship, and other answers into holes, and Hollerith’s electrical readers, counters, and sorters processed the cards. People became machine-readable data, while the state still chose the categories and human punching and sorting labour remained essential.
- Pause and ask
- When tens of millions of lives become holes in cards, what can be counted faster—and what disappears outside the categories?
- How thinking changed
- Encoding responses as holes at predefined positions and reading them through electrical contacts turned human attributes into standardized records a machine could count and sort. An algorithm now involved data format, input labour, classification rules, and machine processing as well as arithmetic.
- What we cannot claim
- The machine did not automate the census as a whole. People visited, recorded, punched, and checked, while the state defined the categories. The 1890 speedup is not credited to Hollerith alone or to one device in isolation.
- This place
- Washington's federal census office concentrated a constitutionally recurring national count, a large budget, and pressure from the previous census backlog. State administration created a market for machines while wielding power to define categories such as race, sex, and citizenship. (Potomac banks · broadleaf trees · low hills · 38.9°N 77.0°W · landmark: United States Capitol (1866))
- Figure board
- Pins pass only through holes punched at predefined positions, closing circuits that move counter dials one step (+1).
- On the river
- A place of ongoing work, marked only by the route’s emblem · 1750–1899 (Cast-iron plaque · gas lamps · iron truss bridge · factory chimneys, railway and steam train · steamboats)
08·1928 CE·Göttingen(basis: Teaching / position)
Hilbert and Ackermann — Can Every Logical Problem Be Decided?
Principles of Mathematical Logic asked whether a finite procedure could determine whether any logical expression is universally valid. The Entscheidungsproblem posed an ambitious question before “mechanical method” had a precise definition. Göttingen marks the Hilbert school’s teaching and research context, not an exact writing room or publication city.
- Pause and ask
- Is there a finite rule-governed procedure that always decides whether any logical expression is universally valid?
- How thinking changed
- Hilbert and Ackermann turned a hope for mechanical solution into the explicit universal *Entscheidungsproblem*. Instead of first specifying a device, they asked whether a decision procedure could halt with yes or no on every input, setting a target later models of computation had to answer.
- What we cannot claim
- The decision problem is not a vague question about making all mathematics easy for people; it asks precisely for a finite procedure deciding universal validity in first-order logic. The 1928 question, 1931 incompleteness, and 1936 uncomputability are connected but not the same theorem.
- This place
- Göttingen's mathematical institute, teaching, axiomatization program, and international student network made foundational questions in logic, geometry, and arithmetic a shared agenda. The pin marks the Hilbert school's research and teaching environment, not an exact writing room or publisher's address. (Wooded hills · broadleaf trees · small river · 51.5°N 9.9°E · landmark: Göttingen Observatory (1816))
- Figure board
- Any logical formula goes in: is there a box that halts after finitely many steps and answers yes (✓) or no (✗)?
- On the river
- A lectern and a board · 1900–1969 (Concrete marker · electric streetlights · concrete bridge · pylons, apartment blocks and radio masts · barges · aircraft)
09·1931 CE·Vienna(basis: Main activity)
Gödel — Formalization Reveals Its Own Limits
Gödel showed that effectively axiomatized systems strong enough for natural-number arithmetic contain statements they cannot decide under specified consistency assumptions, and generally cannot prove their own consistency by their own means. This does not say that all mathematics is wrong or settle the power of every machine not yet defined.
- Pause and ask
- If one formal system is asked to prove every arithmetical truth, what remains once the system can encode statements about itself?
- How thinking changed
- Gödel encoded formulas and proofs as natural numbers, enabling arithmetic to speak about its own provability. Under specified consistency conditions, any sufficiently strong effectively axiomatized system has true statements unprovable within it and generally cannot prove its own consistency internally.
- What we cannot claim
- The incompleteness theorems do not automatically say all mathematics is false or incomplete, that the human mind exceeds every machine, or that computers cannot think. Their assumptions about strength, effective axiomatization, and consistency remain explicit.
- This place
- The University of Vienna, its mathematical colloquium, and discussions of logic and language around the Vienna Circle gave Gödel a community in which to master and debate the Hilbert program's formal tools. Political violence and antisemitism soon fractured that scholarly network. (Danube banks · broadleaf trees · Vienna Woods · 48.2°N 16.4°E · landmark: St. Stephen’s Cathedral (1433))
- Figure board
- Symbols become numbers bundled as 2ᵃ·3ᵇ·5ᶜ; a sentence inside arithmetic then points at its own provability and stays undecided.
- On the river
- A place of ongoing work, marked only by the route’s emblem · 1900–1969 (Concrete marker · electric streetlights · concrete bridge · pylons, apartment blocks and radio masts · barges · aircraft)
10·1936 CE·Princeton(basis: Composition)
Church — Defining Computation Through Function Substitution
Alonzo Church used lambda calculus and recursive functions to formalize effective calculability and gave a negative answer to the general decision problem. At nearly the same time Turing reached the same boundary with a very different machine model. Equivalence among formal models is a theorem; the Church–Turing thesis about every intuitive effective procedure is a distinct claim.
- Pause and ask
- If effective calculation is defined through functions and substitution rules, can the existence of a procedure for every logical problem be settled?
- How thinking changed
- Church established a formal class of computable functions through lambda definability and recursive functions, then showed that no function in that class solves the general decision problem. Turning the intuitive phrase 'mechanical method' into a manipulable mathematical object made uncomputability itself provable.
- What we cannot claim
- That lambda calculus and Turing machines express the same computable functions is a formal result. The Church–Turing thesis identifying every intuitive effective procedure with that class is conceptual and evidential, not another theorem inside one formal system.
- This place
- Princeton's logic seminars, the *Annals of Mathematics*, and a concentration of visitors and graduate students connected the formal work of Church, Kleene, and Rosser with Turing's later study there. The place enabled comparison among models rather than constituting one person's lone school. (Woods · broadleaf trees · gentle lowland · 40.4°N 74.7°W)
- Figure board
- Computation is defined by substitution in lambda terms; the function D that would solve the decision problem lies outside that whole class.
- On the river
- A desk holding a written record · 1900–1969 (Concrete marker · electric streetlights · concrete bridge · pylons, apartment blocks and radio masts · barges · aircraft)
11·1936 CE·Cambridge(basis: Composition)
Turing — Modelling a Procedure as a Machine
Alongside Church’s lambda calculus, Turing modelled a person following symbol-manipulation rules as an abstract machine. The agreement of distinct models helped establish a mathematical account of effective procedure.
- Pause and ask
- By imagining a person who sees one square of paper and follows fixed rules, can the act of computation be defined through minimal operations?
- How thinking changed
- Turing abstracted rule-governed symbol work into an unbounded tape, a reading and writing head, finitely many states, and a transition table. In the same framework he described a universal machine that interprets encodings of other machines and a boundary on deciding in general whether machines halt.
- What we cannot claim
- A Turing machine was not an electronic computer built in 1936 but an abstract model for computability. Turing completeness does not mean fast or intelligent, and the halting result does not prevent analysis of termination for particular restricted programs.
- This place
- Mathematical training at King's College Cambridge, Max Newman's foundations lectures, and access to libraries and journals enabled Turing to derive a formal model from a calculator's physical actions. During review and publication he learned of Church's independent result, leading onward to Princeton. (River Cam · willows · flat fen edge · 52.2°N 0.1°E)
- Figure board
- A head reading one square of an unbounded tape, finitely many states, and one row of a transition table decide what to write and where to move.
- On the river
- A desk holding a written record · 1900–1969 (Concrete marker · electric streetlights · concrete bridge · pylons, apartment blocks and radio masts · barges · aircraft)
12·1937 CE·Cambridge, MA(basis: Composition)
Shannon — Turning Logical Expressions into Relay Circuits
MIT graduate student Claude Shannon connected open and closed relays with Boolean true and false, allowing complex switching circuits to be analysed and simplified as logical expressions. His 1937 thesis is distinct from its 1938 publication, and Victor Shestakov and others produced independent related work. The achievement was a design bridge, not a literal identity of logic and electricity.
- Pause and ask
- Can a circuit of hundreds of relays be analysed and redesigned more simply through logical expressions instead of tracing every wire?
- How thinking changed
- Shannon mapped open and closed relay contacts to Boolean zero and one conditions, writing circuits as expressions and simplifying them by equivalence transformations. A two-way design translation between symbolic logic and physical switches turned circuits from trial-and-error wiring into calculable structures.
- What we cannot claim
- This was not discovery of a natural identity between logic and electricity but a model of specified relay conditions in Boolean expressions. The 1937 thesis is distinguished from its 1938 publication, and independent prior or parallel work by Shestakov and others is not erased.
- This place
- MIT electrical engineering and maintenance work on Vannevar Bush's Differential Analyzer gave Shannon a concrete problem in complex relay circuits, while academic and industrial networks circulated the 1938 paper. Shestakov and others produced independent related work elsewhere in the same period. (Charles River · broadleaf trees · flat land · 42.4°N 71.1°W · landmark: MIT Great Dome (1916))
- Figure board
- A relay circuit, contact x followed by x and y side by side, is calculated like an expression and equals a single-contact circuit (=).
- On the river
- A desk holding a written record · 1900–1969 (Concrete marker · electric streetlights · concrete bridge · pylons, apartment blocks and radio masts · barges · aircraft)
13·1941 CE·Berlin(basis: Main activity)
The Z3 — Instructions on Film Control a Working Machine
Konrad Zuse’s Z3 combined binary floating-point arithmetic, relay control, and instructions on punched film in a working program-controlled machine used for aeronautical calculation. The original was destroyed in the war and familiar museum machines are later reconstructions. Instructions were not stored in internal memory, so precise features matter more than ranking it as the first computer in every sense.
- Pause and ask
- If instructions are punched on film and relays calculate binary numbers, does a paper design become a repeatedly running physical program?
- How thinking changed
- The Z3 read sequential instructions from punched film and performed binary floating-point operations with relays. Separating instructions from hardware let the same machine follow different calculation sequences, while its operational design lacked internal stored programs and general conditional branching.
- What we cannot claim
- The Z3 was a working program-controlled digital calculator, but it did not store instructions in internal memory or demonstrate modern operational generality during the war. Precise capabilities replace a claim to be the world's first computer in every sense.
- This place
- Berlin engineering education, a workshop in Zuse's parents' flat, limited aeronautical-research support, and demand for aircraft calculation linked a largely private construction effort to wartime industry. Bombing destroyed the original and some records; the familiar Z3 is a later reconstruction. (Spree banks · pines · flat land · 52.5°N 13.4°E · landmark: Berlin Cathedral (1905))
- Figure board
- Instruction rows punched on film are read in order 1, 2, 3…, and relays compute binary floating-point numbers ± m · 2ᵉ.
- On the river
- A place of ongoing work, marked only by the route’s emblem · 1900–1969 (Concrete marker · electric streetlights · concrete bridge · pylons, apartment blocks and radio masts · barges · aircraft)
14·1946 CE·Philadelphia(basis: Main activity)
ENIAC — Programming Was Wiring and Human Work
Six women selected from the human computers who calculated firing tables learned ENIAC’s logic and programmed it by arranging cables and switches. Electronic calculation was fast, but changing problems took lengthy setup. This friction and the ENIAC–EDVAC team’s discussions helped develop stored-program design; the history cannot be reduced to one builder or to von Neumann alone.
- Pause and ask
- On an electronic machine without instructions in memory, whose labour—and what kind of labour—was the program?
- How thinking changed
- An ENIAC program decomposed a problem across accumulators, function tables, and control units, then physically reconfigured cables and switches. Electronics made arithmetic dramatically faster while problem changes remained slow, making the value of storing and editing instructions like data newly visible.
- What we cannot claim
- 1946 marks the public unveiling; the machine had already run calculations in 1945. Even where ENIAC is classed as the first general-purpose electronic computer, that is not the same as stored-program operation, and no one group of designers, patrons, builders, or programmers receives the whole achievement.
- This place
- Philadelphia's Moore School assembled U.S. Army Ballistic Research Laboratory funding, electronic engineers, and women who had computed firing tables by hand. Six early programmers learned diagrams and machine logic to make ENIAC work, yet were long marginalized in captions and credit. (Schuylkill banks · broadleaf trees · flat land · 40.0°N 75.2°W · landmark: Philadelphia City Hall (1894))
- Figure board
- Cables and switches physically join accumulator, function-table and control panels: the wiring itself is the program.
- On the river
- A place of ongoing work, marked only by the route’s emblem · 1900–1969 (Concrete marker · electric streetlights · concrete bridge · pylons, apartment blocks and radio masts · barges · aircraft)
15·1948 CE·Manchester(basis: Main activity)
Manchester Baby — Writing Instructions into Memory
On 21 June 1948 the Small-Scale Experimental Machine, or Baby, ran a seventeen-instruction highest-factor program stored in electronic memory. Freddie Williams, Tom Kilburn, Geoff Tootill, and expertise and equipment from the wartime Telecommunications Research Establishment made the experiment possible. The important shift is the editable stored program, not an absolute first-computer ranking.
- Pause and ask
- If instructions occupy the same memory as numbers, can a machine's task change without rewiring its cables?
- How thinking changed
- The Manchester Baby fetched instructions from Williams–Kilburn electronic memory, repeated according to results, and ran a seventeen-instruction highest-factor program. As a program became a readable and editable memory pattern rather than fixed wiring, machine construction and program writing began to separate operationally.
- What we cannot claim
- The Baby was a small experimental memory testbed, not claimed as the first computer in every category. Its key evidence is an executed electronically stored program, with joint credit to Williams, Kilburn, Tootill, and the wartime research network rather than a ranking slogan.
- This place
- The University of Manchester brought together radar-memory expertise carried from the wartime Telecommunications Research Establishment by Williams, Kilburn, and Tootill, university laboratory space, and a later Ferranti industrial link to test an unstable new memory in a small experimental machine. (River Irwell · broadleaf trees · flat basin · 53.5°N 2.2°W · landmark: Manchester Town Hall (1877))
- Figure board
- Seventeen instructions and the numbers share one electronic memory; instructions are fetched in turn and the run loops back on results.
- On the river
- A place of ongoing work, marked only by the route’s emblem · 1900–1969 (Concrete marker · electric streetlights · concrete bridge · pylons, apartment blocks and radio masts · barges · aircraft)
16·1956 CE·Amsterdam(basis: Main activity)
Dijkstra — From Machine Demonstration to Reusable Algorithm
For the 1956 inauguration of the ARMAC computer, Dijkstra designed a method for finding the shortest route between Dutch cities so a general audience could understand the demonstration. He later recalled conceiving it over coffee with his fiancée Ria in about twenty minutes, but publication came in 1959. A reusable algorithm also needs its conditions made explicit: this method cannot simply accept negative edge weights.
- Pause and ask
- What must be stated to turn a shortest-route idea into a procedure whose answer can be checked on other maps and computers?
- How thinking changed
- Dijkstra's method extends settled shortest distances from a source by repeatedly selecting the nearest unsettled vertex. A one-off demonstration problem became a reusable algorithm with an input graph, selection rule, termination, and correctness condition—and a visible boundary where negative edge weights break the guarantee.
- What we cannot claim
- The twenty-minute café story is Dijkstra's later recollection, not a contemporaneous record. The method was devised for a 1956 demonstration but published in 1959, and it cannot be applied unchanged to graphs with negative edge weights.
- This place
- Amsterdam's Mathematisch Centrum ARMAC team needed a demonstration whose result non-specialists could understand when inaugurating a Dutch-built computer. A legible city map and the real machine's limits favoured an elegant, memory-conscious procedure. (Amstel and canals · elms · flat polders · 52.4°N 4.9°E · landmark: Westerkerk (1638))
- Figure board
- On a weighted graph the nearest unsettled vertex is settled one at a time, growing a shortest-path tree, under the condition w ≥ 0.
- On the river
- A place of ongoing work, marked only by the route’s emblem · 1900–1969 (Concrete marker · electric streetlights · concrete bridge · pylons, apartment blocks and radio masts · barges · aircraft)