Classicery · 004B · 2026
The fourth room in the theory wing — the one that starts at the bottom. Before a machine can fill a gap or make what you never drew, somebody had to work out how to tell it anything at all. That telling is the program. One axis runs the whole room: how much you spell out, and how much the machine has agreed to supply.
“Paragraphs on Conceptual Art,” Artforum, 1967 · the wall drawings
Before the first machine, a program that ran on people. LeWitt signed the instruction, not the drawing — a paragraph precise enough that any crew of draftsmen could execute it on a wall. The oldest definition of a program there is, written in plain English, decades before code.
Notes on Menabrea’s sketch of the Analytical Engine, 1843 · Note G
The first program, written for a machine never built. Note G lays out a full procedure for computing the Bernoulli numbers — usually counted as the first published program. The Engine, she wrote, weaves algebraic patterns as the Jacquard loom weaves flowers and leaves; and it has no pretensions to originate anything. It only does what we order it to do.
“On Computable Numbers, with an Application to the Entscheidungsproblem,” 1936
Not what one machine could be told to do, but whether one machine could be told to be any machine. Turing’s universal machine reads a description off its tape and becomes whatever the description says. The program stops being a procedure for a particular engine and becomes a piece of writing a general machine reads. Remember him — he returns at the far end of the wing.
“First Draft of a Report on the EDVAC,” 1945 (von Neumann; the single byline slighted Eckert & Mauchly)
Where does the description live? The stored-program idea says: put the instructions into memory, in the same form as the data. The program becomes numbers the machine merely holds — loadable, replaceable, savable, copyable, able to read and write other programs. The description stops being the wiring. It becomes, for the first time, software.
ENIAC, 1945 · Jean Jennings Bartik, Betty Holberton & team
The cards and wires you asked for. To program ENIAC you rewired the machine — switches set by hand, cables run across a wall of plugboards. Six women, handed a machine no one had ever programmed, invented the practice as they went, and were left out of the photographs for fifty years. The bottom rung: the human specifies every pulse, the machine supplies nothing.
The A-0 system, 1952 (nearer a loader/linker than a modern compiler)
The keystone of the room, and almost nobody believed it. Hopper proposed that people write in something like English and let a program translate it down to machine code. She had it running in 1952. The reason they gave for refusing it: a computer could only do arithmetic. The Canon4 thesis, stated seventy years before it fully arrived.
FORTRAN, IBM, 1957 · the deck, the batch, the one typo that kills the run
Hopper’s idea made respectable, by winning an argument about speed. The fear was that a machine translating your formulas would be too slow to allow, so Backus’s team made the compiler nearly as fast as a careful human. And the object of the era: the deck of punch cards, one line per card, where a single mistyped character killed the run and the loop from idea to result was measured in hours.
Dartmouth, 1964 · built on a time-sharing system, for every student
The first abstraction aimed, on purpose, at non-programmers. Kemeny and Kurtz built a language deliberately easy and a system that answered in seconds, so a student who was not an engineer could write a working program in an afternoon. The question stopped being how experts work better and became who is allowed to instruct the machine at all. Their answer was everyone.
VisiCalc, 1979 · Apple II · the most-used programming model on earth
The most widely used programming language in the world, and almost no one who uses it believes they are programming. A cell holds a formula that refers to other cells; you declare the relationships and the sheet recomputes itself, downstream, on its own. That is dataflow — specify what relates to what, not when to compute. Remember the grid that recalculates itself. It comes back.
Design By Numbers (MIT Press, 1999) · the Aesthetics + Computation Group
The on-ramp built by a designer, on purpose, as an argument: that the medium is made of code and the people who shape how things look should be shaping it. Out of his group came Processing, which taught a generation of designers to write their own work. The instruction stops being something done to design by others, and gets claimed as part of the craft.
Sketchpad 1963 · Bert Sutherland’s TX-2 dataflow 1966 · GRaIL, RAND, 1968
The other path up the ladder: you do not write the instruction, you draw it. A Sketchpad constraint is an instruction the machine satisfies for you. Bert Sutherland’s 1966 dataflow graph and RAND’s 1968 runnable flowchart take the web of dependencies you keep in your head and pin it to the wall, where you can see it and rearrange it. The playground rooms are next door; this is the wall text they never had.
Max, IRCAM, ~1988 (after Max Mathews) · MSP audio layer, 1997
The graph left the lab and found the artists first, at a music institute in Paris. In Max you place boxes and run patch cords between them, and the signal flows down the wires you laid. Not writing music in the old sense, not writing code — building a little machine that makes sound from the relationships you declared. Dataflow grew up and became something you perform with.
Generative Components, Aish/Bentley (2003–2007) · Grasshopper, Rutten/McNeel, 2007 (“Explicit History”)
The graph reaches the designer’s desk. Wire components together, move a slider, watch the geometry rebuild down the chain — the spreadsheet’s engine wearing new clothes. It is gorgeous for relationships and clumsy for control, which is why a serious user reaches for a scripting box. Rutten built it to spare designers from code, then was startled to find it taught “at gunpoint” in universities. The relief became a rigor.
Classicery · 004B · the door to Volume Five
The top rung is the one Hopper described in 1952 and was told was impossible: say what you want, in plain words, and the machine supplies the form. Here the instruction room ends, because here it stops being only about instruction — specifying intent and letting the machine make what you never drew turn out to be one motion. The arc of telling and the arc of making meet on this threshold. Walk through.
Instruction is the fourth room in Classicery’s theory wing — From the Literature, Volume Four. Volume One was the surface, Volume Two the page, Volume Three the motion. This one is the foundation none of them built: what a program even is, and how the act of telling a machine what to do became, rung by rung, more human.
Fifteen tracks, from Ada Lovelace’s Note G in 1843 to the moment you can simply ask in plain words. One axis runs the whole room — how much you specify, and how much the machine has agreed to supply — and a ladder climbs it: cards and wires, the first abstractions, languages for the rest of us, the drawn graph, and the natural-language hinge that opens onto Volume Five.
The room is built to feel like the ladder it describes. Its lattice begins rigid and snapped, and loosens as the playhead climbs — by the hinge the grid has all but dissolved. Turn it off with your system’s reduce motion setting and it settles into a plain, legible reading room.
The wing’s refrain — that works fine in practice, but what about the theory? — is an orphan. Trace it and it leads nowhere clean: it has been pinned on a Chicago economist, an Irish statesman, a South Carolina senator, a Heidelberg professor, and no one in particular, with the earliest strong sighting in a magazine in 1911 that named none of them. It is quoted so freely that no one knows who said it first — which is the joke’s last turn, and why it belongs over this wing. A line that works fine in practice, with no settled theory of where it came from. Its origin remains, fittingly, a question of theory.
Drafted by Claude with Phil Renato, 2026. No invented quotations from any real person, living or dead; living authors are paraphrased and quoted only from the record. Every date and attribution passed a two-pass fact-check. Audio is read by a single narrator; tracks without audio yet show a PENDING state. CART 004B / 027 · THEORY WING · FROM THE LITERATURE.