Microprocessor 1971 Calculator Chip Became General Purpose Processor
# Why a Calculator Chip Project Became a General-Purpose Processor: The First Series Whose Silicon Does Not Fix What the Chip Does
## 1. Why Every Earlier Series Encoded Its Function Permanently, and This One Does Not
Busicom asked Intel to build a set of custom chips for a desktop calculator, following the same approach this project's own earlier series have always used — a circuit whose function is fixed the moment its mask set is drawn, so that one chip is permanently an adder, another permanently a register, another permanently a display driver. Ted Hoff's proposal at Intel replaced that entire approach with something this project has never built before: a single chip that implements a general-purpose interpreter, reading a sequence of instructions from memory and carrying out whatever that sequence specifies, so that what the chip actually does on any given day is decided by the program loaded into it, not by the silicon itself. Every series this project has documented up to this one built hardware whose behavior was permanent the instant fabrication finished. This series builds the first chip whose silicon is finished and permanent, but whose behavior is not.
where $F_{\text{silicon}}$ is the fixed function a chip's own fabricated structure permanently implements — true of every logic gate and every memory cell this project has built before now — while $F_{\text{processor}}$ is itself a fixed, general interpreter circuit, built once, whose output behavior depends on $\text{Program}(t)$, a sequence of instructions read from separate memory that can change completely without changing a single transistor on the processor chip itself.
## 2. Real Diagram: Three Chips Where Earlier Proposals Needed Many More
The sketch below previews the structure Hoff's proposal actually required — a general processor, a separate chip to hold the program, and a separate chip to hold working data — in place of the larger set of dedicated, fixed-function chips a calculator built in this project's earlier style would have needed.
## 3. The Same Idea 1968 Established, Carried to a New Scale and Purpose
The 1968 DRAM series opened this project's own memory thread with a single requirement: a device that remembers a bit of information across time, something no logic gate this project had built before it could do. This step's processor needs memory for a related but genuinely different reason. A DRAM cell remembers one bit; this series' processor needs memory that remembers an entire ordered sequence of instructions, and needs to consult that sequence not once, but at every single clock cycle, to decide what to do next. The persistence 1968 introduced as a device-level property becomes, in this series, the entire basis for how a chip's function gets defined — not fixed permanently in silicon the way every earlier series' own circuits were, but read continuously from a memory this series will build using exactly the kind of device 1968's own series first characterized.
Step 1 does not propose a faster or cheaper calculator chip; it proposes a chip whose function this project has, for the first time, separated entirely from its own fabrication, carrying 1968's own insight about persistence into a role no single bit of stored charge was ever asked to play before.