Material Culture of Computation by Dr Doron Swade

Rina Chen’s living notebook on digital craft and design.

![[Pasted image 20260215140536.png]]

  • material culture/object history/functions/shapes used to be distinct
  • converged in the 70s, allowed by submicron transistors/silicon

  • the death of material culture

  • Babbage: difference engine, first machine to embody mathematical rule in mechanism, digital automation
    • Lady Byron commented: A thinking machine
    • but algorithm is executed by human
    • Context: industrialization of thought
    • a digital attempt 1820:
  • ![[Pasted image 20260215144257.png]]
  • discrete, incremental fixed jumps ![[Pasted image 20260215144507.png]] ![[Pasted image 20260215144651.png]]
  • Analytical engine in 1833 marked the transition from calculation to computation, generalization
    • calculation: rules cannot be altered
    • punch card: algorithm is transferred to operation card, programmable ![[Pasted image 20260215151328.png]] ![[Pasted image 20260215151422.png]]
  • specific design features
    • separation of memory and processor
    • microprogramming: small program to solve larger program
    • if or
  • Jacquard loom and punch card
    • exquisite subtlety is not the exclusive province of a human agent
    • to improve acceptance of industrial objects
    • specificity is in the card, not the machine, software is separated from hardware
  • From mathematical machine to general purpose computer
    • the realization the the number can represent entity other than quantity ![[Pasted image 20260215152605.png]]
    • power of computing comes from the representational power of numbers
    • we can manipulate symbols to produce knowledge of the world ![[Pasted image 20260215153004.png]]

      Reference

Dr Doron Swade, lecture in Gresham College

More History

1. Purpose of the Series

  • Explains computer science as a field, not programming tutorials.

  • Covers the journey from:

    • Bits & transistors

    • Logic gates & CPUs

    • Operating Systems

    • Internet, VR, Robots

  • Focus: how computing works and its societal impact.

2. Importance of Computers in Modern Life

  • Computers run critical infrastructure:

    • Power grids

    • Transportation

    • Finance

    • Manufacturing

    • Communication

  • Compared to the Industrial Revolution in terms of societal transformation.

  • Current era described as the Electronic Age.

3. Layers of Abstraction

  • Computers are simple machines at the lowest level but appear complex because of many abstraction layers.

  • Users and engineers typically interact with higher layers, not raw hardware.

  • Goal: understand computing from 1s and 0s → full systems.

4. Origins of Computing

Abacus (~2500 BCE, Mesopotamia)

  • First widely recognized computing device.

  • Manual calculator for addition/subtraction.

  • Stored intermediate results like modern memory.

  • Used place-value rows (units, tens, hundreds, etc.).

5. Other Early Computing Tools

Over ~4000 years humans built devices to reduce mental effort:

  • Astrolabe – navigation and latitude at sea.

  • Slide Rule – multiplication/division.

  • Mechanical clocks – time and astronomical calculations.

Theme: Tools amplify human intelligence and reduce labor.

6. “Computer” Originally Meant a Person

  • First recorded use (1613) referred to humans who performed calculations.

  • Term shifted to machines in the late 1800s.

7. Mechanical Calculators

Step Reckoner (1694)

  • Built by Gottfried Wilhelm Leibniz.

  • Gear-based machine performing:

    • Addition

    • Subtraction

    • Multiplication

    • Division

  • Influenced calculator design for ~300 years.

8. Pre-Computed Tables

  • Before modern computers, people used large books of:

    • Square roots

    • Log tables

    • Artillery range tables

  • Saved time but were slow to update and error-prone.

9. Military Need for Computation

  • Artillery firing required accounting for:

    • Distance

    • Wind

    • Temperature

    • Air pressure

  • Range tables were effective but inflexible.

10. Charles Babbage and Early Computer Concepts

Charles Babbage

Often called the “Father of Computing.”

Difference Engine (1820s)

  • Mechanical machine to automatically compute polynomial tables.

  • Never completed in his lifetime.

  • Successfully built from his plans in 1991.

Analytical Engine

![[Pasted image 20260308174143.png]] (from cs.stanford.edu)

  • Concept of a general-purpose computer.

  • Included:

    • Memory

    • Sequential operations

    • Input/output

    • Printer

  • Not built, but introduced the idea of programmability.

11. First Programmer

Ada Lovelace

  • Wrote theoretical programs for the Analytical Engine.

  • Recognized computing as a new symbolic language.

  • Considered the first computer programmer.

12. Census and Punch Cards

Herman Hollerith

  • Built an electro-mechanical tabulating machine for the 1890 U.S. Census. a punch card is placed on pools of mercury and a metal spring would be able to penetrate the punchcard to touch mercury if there’s a hole, forming an electrical contact, used electromechanical solenoids to increment mechanical counters btw, flipdot = solenoid(Electromagnetism) + Magnetic dipole interaction + Mechanical rotation

  • Used punch cards to encode data.

  • Achieved ~10× speed improvement.

  • Reduced census processing from ~13 years to ~2.5 years.

13. Birth of IBM

IBM

  • Originated from Hollerith’s Tabulating Machine Company (1924 merger).

  • Became a dominant force in business computing.

14. Transition to Digital Computers

  • Electro-mechanical machines transformed government and commerce.

  • Rapid population growth and globalization demanded faster, more flexible systems.

  • Set the stage for fully electronic digital computers.

1. Growing Complexity in the Early 20th Century

  • Rapid increases in:

    • World population

    • Global trade

    • Scientific and engineering projects

    • Military mobilization (World Wars I & II)

  • Result: Huge demand for faster, automated computation.

  • Electro-mechanical machines became:

    • Room-sized

    • Expensive

    • Error-prone

    • Slow

2. Electro-Mechanical Computers

IBM Harvard Mark I (1944)

  • One of the largest electro-mechanical computers.

  • ~765,000 components, 500 miles of wire.

  • Used for WWII simulations (e.g., Manhattan Project).

  • Relied on relays as switches.

  • Performance:

    • 3 additions/subtractions per second

    • Multiplication: ~6 seconds

    • Division: ~15 seconds

  • Problems:

    • Slow switching speed

    • Mechanical wear and frequent failures

    • Required constant maintenance

3. Relays and “Bugs”

  • Relay = electrically controlled mechanical switch.

  • Limited speed (~50 switches per second).

  • Mechanical parts wore out easily.

  • In 1947, Grace Hopper documented a moth stuck in a relay → origin of the term “computer bug.”

4. Vacuum Tubes – First Electronic Switches

John Ambrose Fleming (1904)

  • Invented the vacuum tube diode.

  • Allowed one-way current flow.

Lee de Forest (1906)

  • Added a control electrode → triode vacuum tube.

  • Acted like an electronic switch.

  • Advantages over relays:

    • No moving parts

    • Thousands of switches per second

  • Disadvantages:

    • Fragile

    • Generated heat

    • Burned out like light bulbs

    • Expensive initially

5. First Electronic Computers

Bletchley Park

Colossus (1943)

  • Designed by Tommy Flowers.

  • First large-scale programmable electronic computer.

  • Used ~1,600 vacuum tubes.

  • Purpose: decrypt Nazi communications.

  • Programming done via plugboards (manual wiring).

ENIAC

  • Built by John Mauchly and J Presper Eckert.

  • First general-purpose programmable electronic computer.

  • Speed:

    • ~5,000 additions per second.

  • Limitations:

    • Frequent vacuum tube failures.

    • Operated only ~12 hours before breakdowns.

6. The Transistor Revolution (1947)

Inventors at Bell Labs

  • John Bardeen

  • Walter Brattain

  • William Shockley

Key Advantages:

  • Solid-state (no fragile glass)

  • Much smaller

  • Faster switching (10,000+ per second initially)

  • Lower power consumption

  • More reliable

  • Cheaper over time

7. Early Transistor Computers

IBM IBM 608 (1957)

  • First fully transistorized commercial computer.

  • ~3,000 transistors.

  • Thousands of operations per second.

  • Marked transition from government-only machines to business and home computing.

8. Silicon Valley and Semiconductors

  • Region between San Francisco and San Jose became Silicon Valley.

  • Named after silicon, key semiconductor material.

  • Company lineage:

    • Shockley Semiconductor

    • Fairchild Semiconductor

    • Intel

9. Scale and Speed Today

  • Modern transistors:

    • < 50 nanometers in size.

    • Switch millions to billions of times per second.

    • Can run reliably for decades.

  • Enabled the miniaturization of computers into phones and laptops.

Core Takeaway

Computing hardware evolved through three major switching technologies:

Relays → Vacuum Tubes → Transistors

Each step dramatically improved speed, reliability, size, and cost, transforming computers from fragile room-sized machines into compact, powerful electronic devices and laying the foundation for modern digital computing.

1. Ladder of Abstraction

  • Computing is built in layers of abstraction.

  • Lower levels: physical switches, gears, transistors.

  • Higher levels: logic gates → circuits → processors → software.

  • Engineers and programmers usually work at higher layers, not at the raw hardware level.

2. Binary Representation

  • Computers use Binary = two states.

  • Electrical states:

    • On (current flowing) = True = 1

    • Off (no current) = False = 0

  • Two states are preferred because:

    • More reliable and resistant to electrical noise.

    • Easier to distinguish than 3+ states.

    • Supported by existing mathematics (Boolean Algebra).

3. Boolean Algebra

  • Mathematical system dealing with True / False values.

  • Created by George Boole in the 1800s.

  • Unlike regular algebra (numbers), Boolean algebra uses logical operations.

Three Fundamental Operations

  1. NOT – flips value

    • True → False

    • False → True

  2. AND – both inputs must be true.

  3. OR – at least one input must be true.

4. Transistors as Switches

  • A transistor acts like an electrically controlled switch.

  • Has:

    • Control wire (input)

    • Two electrodes (current path)

  • Turning control on/off allows or blocks current.

  • Logical behavior can be built physically using transistors.

5. Logic Gates

Logic gates are physical circuits that implement Boolean operations.

NOT Gate

  • One input, one output.

  • Output is the opposite of input.

AND Gate

  • Two inputs.

  • Output is true only if both inputs are true.

  • Built by placing transistors in series.

OR Gate

  • Two inputs.

  • Output is true if either input is true.

  • Built by placing transistors in parallel.

6. XOR (Exclusive OR)

  • Special logic gate.

  • Output is true only when inputs are different.

  • False when both are true or both are false.

  • Built using combinations of NOT, AND, and OR gates.

  • Very important for arithmetic and processor design.

7. Abstraction in Engineering

  • Engineers rarely design at the transistor level.

  • They use:

    • Logic gates

    • Larger components made from gates

    • Eventually CPUs and software.

  • Each layer hides the complexity below it.

8. Computation from Logic

  • Even simple gates can evaluate complex logical statements.

  • True/False signals become the first form of data representation.

  • From these basics, full computers are built.

1. Everything Starts with Binary

  • Computers use binary (base-2) instead of decimal (base-10).

  • Binary has only two digits: 0 and 1.

  • These correspond to electrical states:

    • 1 = on / true

    • 0 = off / false

  • Larger numbers are represented by adding more binary digits, just like adding more decimal digits.

2. Positional Number Systems

Decimal (Base-10)

  • Each column is a power of 10: ones, tens, hundreds, etc.

Binary (Base-2)

  • Each column is a power of 2:
    1, 2, 4, 8, 16, 32, 64, 128…

  • Example:
    Binary 101 = 1×4 + 0×2 + 1×1 = 5 (decimal)

3. Bits and Bytes

  • Bit = one binary digit (0 or 1).

  • Byte = 8 bits.

  • 8 bits can represent 256 values (2⁸) → range 0–255.

  • Early “8-bit” systems were limited in:

    • Colors

    • Sound variation

    • Numerical range

4. Data Size Units

  • Kilobyte (KB) ≈ 1,000 bytes (or 1,024 in binary context)

  • Megabyte (MB) ≈ 1 million bytes

  • Gigabyte (GB) ≈ 1 billion bytes

  • Terabyte (TB) ≈ 1 trillion bytes

(Binary definition uses powers of 2, decimal uses powers of 10.)

5. Word Size: 32-bit vs 64-bit

“Bitness” = how many bits a computer processes at once.

32-bit

  • Max unsigned value ≈ 4.3 billion

  • Often uses 1 bit for sign → ±2 billion range.

64-bit

  • Max value ≈ 9.2 quintillion

  • Necessary for:

    • Large numbers

    • Huge memory addressing (GBs/TBs of RAM)

6. Representing Negative Numbers

  • Common approach: first bit = sign

    • 0 = positive

    • 1 = negative

  • Remaining bits store magnitude.

7. Floating Point Numbers (Decimals)

Used for non-whole numbers like 3.14 or 12.7.

IEEE 754 Standard

Stores numbers similar to scientific notation:

  • Sign bit – positive or negative

  • Exponent – scale (power of 10 or 2)

  • Significand (Mantissa) – actual digits

Example concept:
625.9 → 0.6259 × 10³

8. Representing Text with Numbers

ASCII (1963)

  • 7-bit code → 128 characters

  • Includes:

    • A–Z, a–z

    • Digits 0–9

    • Punctuation

    • Control characters (like newline)

  • Enabled interoperability between different computers.

  • Limitation: mostly English-only.

9. Character Encoding Problems

  • Countries extended ASCII differently.

  • Opening foreign text often produced garbled symbols.

  • In Japan this issue was called “mojibake” (scrambled text).

10. Unicode – Universal Text Encoding (1992)

  • Designed to replace all national encodings.

  • Supports 100,000+ characters from many languages.

  • Includes:

    • Global scripts

    • Math symbols

    • Emoji

  • Common implementations use 16 bits or more, allowing over a million possible codes.

11. Beyond Numbers and Text

Other file formats also use binary:

  • Images – pixel colors

  • Audio (MP3) – sound waves

  • Video – sequences of images + sound

  • Operating Systems & Programs – all binary internally

Big Idea

Computers don’t just store numbers — their real power is computation: manipulating numbers in structured ways (add, subtract, compare, etc.).
The component that performs this work is the ALU — Arithmetic Logic Unit, often called the mathematical brain of the computer.

What the ALU Does

The ALU has two halves:

1. Arithmetic Unit

Handles number math:

  • Addition

  • Subtraction

  • Incrementing (add 1)

  • Sometimes multiplication/division (in advanced CPUs)

2. Logic Unit

Handles logical operations:

  • AND

  • OR

  • NOT

  • XOR

  • Tests like “is this number zero?” or “is it negative?”

Building Addition from Logic Gates

Half Adder (1-bit addition)

Adds two single bits A and B.

Outputs:

  • Sum → produced by XOR

  • Carry → produced by AND

Why carry?
Because 1 + 1 = 10 in binary. The 1 moves to the next column.

Full Adder (3-bit addition)

Adds:

  • Bit A

  • Bit B

  • Carry-in from previous column

Outputs:

  • Sum

  • Carry-out

A full adder is made from:

  • Two Half Adders

  • One OR Gate

Multi-Bit Addition: Ripple Carry Adder

To add 8-bit numbers:

  1. First column uses a Half Adder.

  2. All other columns use Full Adders.

  3. Carry bits “ripple” forward from right to left.

Overflow

If the final carry exceeds the available bits, the number is too large → Overflow error.

Classic example: early arcade games like Pac‑Man broke after level 255 because they only stored levels in 8 bits.

Speed Consideration

Ripple carry is simple but slightly slow because each carry waits for the previous one.
Modern CPUs often use Carry-Lookahead Adders to speed this up.

Multiplication & Division

  • Simple processors do not have dedicated circuits.

  • They perform repeated addition/subtraction.

  • More advanced CPUs include dedicated hardware for speed.

Logic Unit Examples

  • Checking if result equals 0 (Zero Test)

  • Checking if result is negative

  • Bitwise AND/OR/XOR operations

Operation Codes (Opcodes)

The ALU needs instructions telling it what to do.

Example (4-bit control signals):

  • 1000 → Add

  • 1100 → Subtract

These are called operation codes.

Flags (Status Outputs)

After a calculation, the ALU outputs tiny 1-bit signals called flags:

  • Zero Flag → result is 0

  • Negative Flag → result is negative

  • Overflow Flag → number too large

These flags help the CPU make decisions (like comparisons and branching).

Historical Reference

A famous early ALU chip was the Intel 74181:

  • 4-bit only

  • ~70 logic gates

  • Major milestone in miniaturization

Core Insight

An ALU is just a very clever arrangement of logic gates.
From simple TRUE/FALSE switches, computers build:

  • Arithmetic

  • Logic

  • Decision-making

  • Ultimately, all computation.