[Up- imhofaq] [Robot Wisdom home page]

Timeline of knowledge-representation

[I've had to break this up into four subpages so far, linked below.]

"In most cases, the model is a partial representation of the
system, and the quality of its representation is determined

by comparing its behavior with that of the real system."
--Evans et al, Simulation Using Digital Computers 1967

"Everything that rises must converge." --Teilhard de Chardin

"She's got the right dynamic for the New Frontier."
--Donald Fagen (b1948) The Nightfly

The evolution of computer science can be viewed as an evolution of data-representations towards ever-greater comprehensiveness and 'quality'. This timeline will try to reconstruct what classes of data were being processed at each point, starting with the earliest prehistory, and how changes in representation allowed new classes, and new styles of processing.

In general, representations in each area of knowledge will be expected to follow an approximate progression from pre-verbal intuitions to nonverbal images to words and symbols to sorted symbols to counting and measurement to algorithmic relationships to fullscale simulations. (Sometimes this progression strays into decadence, with falsely complicated symbol systems. At other times, word-symbols may be denigrated as inferior to numbers.)

Each of the various domains of knowledge will follow this path at its own pace, including eg physics, chemistry, astronomy, geology, biology, and psychology. Paradoxically, the more-complex psychological domains were usually articulated verbally at a very early period, but the simpler physical domains were the first to be precisely simulated. (Early psych-related sims dealt mainly with economics and war.)

One long-range goal of this survey is the design of an AI-based operating system that includes deeply-integrated 'hooks' for each of these classes of data. [DecentOS]

(Mueller offers these synonyms for 'model': "representation, abstraction or concretion, idea or idealization, illustration, sensualization or view, pattern or scheme, shape and configuration, picture, symbol, sign and icon, metaphor and allegory, example and analogy, fiction and vision, draft and plan, prototype and archetype, paradigm and exemplar, system and hypothesis, theory, philosophy, treatise and principles, doctrine and teachings, law, rule, formula, archetype, copy, design, ectype, example, facsimile, image, imitation, mold, original, pattern, prototype, replica, type, ideal". cite)

I'm starting to explore topical summaries for different domains-- one experiment (miltary sims) is included below.

Jump to main timeline:
page one: start 10000BC 1000BC 500BC 250BC 0 500 1000 1300 1500
page two: 1600 1700 1800 1830 1860 1880 1900 1910 1920 1930
page three: 1938 1942 1946 1950 1954 1957 1959 1961 1964 1967
page four: 1970 1973 1977 1981 1985 1989 1992 1994 1996 1999

The history of knowledge-representation is probably reflected best in the evolution of military simulations. Fighting, defending, and also hunting intelligent prey are vital skills that go back a billion years. Kittens when they wrestle are unconsciously simulating more-serious grownup fights.

The memetic-evolutionary process of 'externalising' these skills in representations may have started with the gestural sign-language of elders, training youths for coordinated group attacks. Gradually this guidance was verbalised, and it became possible to plan group-strategies in advance, and retell war stories at night around the fire.

The shaping of weapons reflected increasing sophistication in anticipating their effectiveness in battle. Training involved a growing body of expertise. Heroes must have been rewarded with special body-ornaments. With the invention of written language c3500BC, leaders quickly took to praising their own deeds on monuments. Treaties were also written out formally.

The first written number-systems allowed for much more efficient inventories of the growing armies' weaponry and provisions. The earliest theoretical treatment of warfare was Sun Tzu's "Art of War" c500BC. The earliest historical treatment was probably Herodotus c440BC, and Greek playwrights explored the psychology and ethics of fighting. (Although Aristotle tutored Alexander the Great, his inventories of human knowledge omitted military matters, except indirectly in the 'Politics' and 'Ethics'.)

Weaponry continued to evolve, with the help of spoken and written language, but it wasn't until 1537 that mathematics became critical, with Tartaglia's invention of the gunner's quadrant, and his publication of the first 'firing tables' for different weapons. Galileo then worked out the mathematics of parabolic ballistics in 1609.

The industrial revolution c1750 introduced increasing mathematical precision into the construction of cannons and other weapons. Simultaneously, miltary training and planning began to use models and games. (By 1898, wargames were popular enough among civilians to inspire the creation of Jane's first reference work, "All the World's Fighting Ships".)

Between the Boer War and World War I, wargames for training and planning began to reveal both the strengths and weaknesses of their limited models. Logistics-models (eg) could suggest more efficient strategies for supplying the front, but simply knowing the size, weaponry, and disposition of the armies was not enough to predict the winner-- various human factors (eg sabotage, diplomacy, and morale) also had to be taken into account.

During WW1, the USA experimented with punchcard technology to manage various resources. The first psychological tests were invented to better fit inductees to their assignments. In 1916, Lanchester tried to use mathematical techniques to analyse air- warfare, and in 1919 the Quaker pacifist Lewis Fry Richardson tried to model the "Mathematical Psychology of War".

German wargames after WW1 introduced experts to more- accurately fill the roles of diplomats, journalists, etc. Flight simulators evolved during the 1930s to train pilots quickly, cheaply, and safely.

Moulton in 1926 introduced much more complicated formulae for ballistics, which demanded vastly improved calculating machines. Vannevar Bush's analog differential analyzer began filling this role in 1932. The British invented 'operations research' c1937, applying mathematical analysis to tactical planning. Experts were assigned during WW2 to search for ways of undermining the enemy's economy.

The German rocket program put Konrad Zuse to work, building some of the first computers for aerodynamic calculations, while the British assigned Alan Turing to break codes. WW2 also saw increasing mathematical sophistication in designing bombs, airplanes, and radar. And the atomic bomb program of course required massive amounts of computation. Simulators were designed to train soldiers in various technical skills. Analog devices including gyroscopes were explored to help aim longrange guns.

After WW2, the prospects of nuclear war inspired the Pentagon to create the RAND thinktank, using wargames and computers to completely rethink strategies and systems for the 'Cold War'. Social scientists at RAND were encouraged to create mathematical models that often bore only a faint resemblance to reality. Von Neumann's game theory required endlessly sophisticated math, but offered only dubious wisdom.

The USA's offense focused on guided missiles, their defense on radar detection of Russian planes and missiles via Jay Forrester's massive SAGE project (begun in 1950). The Pentagon quickly embraced digital computers for day-to-day logistics, so RAND also explored mundane issues like optimising inventories.

Awkward attempts were made to translate Russian scientific literature automatically, and to store it in searchable databanks. Civilian wargames achieved new levels of authenticity during the 1950s and 60s under the leadership of Avalon Hill. RAND tried to introduce computers into wargaming as early as 1954, and slowly but surely these models increased in accuracy and utility.

Sputnik in 1957 turned the missile-race into a spacerace, with new emphasis on basic scientific research at ARPA. JFK and Robert McNamara in 1961 introduced the RAND approach thruout the Pentagon, and LBJ would expand this optimistically to every department. (The Vietnam War eventually demonstrated some realworld limitations of RAND-style modeling.)

Computer models were used to test the designs of complex weapons like tanks, and they became increasingly useful in wargames. By the late 1970s, civilian innovations in wargame-design were being adapted by the Pentagon. 1982 saw the US Air Force exploring AI for mission planning, and the same year RAND wargames in which the computer itself first did the enemy's planning. By 1988, a networked simulation 'SimNet' had been created in which each computer took the role of a single combat vehicle.

Chris Crawford in 1985 offered the geopolitical sim "Balance of Power".

page one: start 10000BC 1000BC 500BC 250BC 0 500 1000 1300 1500
page two: 1600 1700 1800 1830 1860 1880 1900 1910 1920 1930
page three: 1938 1942 1946 1950 1954 1957 1959 1961 1964 1967
page four: 1970 1973 1977 1981 1985 1989 1992 1994 1996 1999

A much earlier version of this timeline was based mostly on "The Brain Makers" by HP Newquist, plus Kurzweil's "The Age of Intelligent Machines" [timeline]. Early calculator-history below is from Eames' "A Computer Perspective"; industrial-revolution stuff via Burke's 'Connections'. Another goldmine: the online archives of "Journal of Artificial Societies and Social Simulation" (JASSS).

[rcl] = Reed C Lawlor "Information Technology and the Law" in Advances in Computers 1962 yearbook

Lehmann's ontologies

Bateman's portal

IT timelines

Greenspun MIT

Columbia

Phasor pdfs

Computational linguistics: SWan

classification schemes

You can submit a new URL or any other suggestion for this page by typing it into the box below. It will instantly become visible to anyone at this comments page. I should get around to checking it out and updating it above within a week or three, at which point I'll delete it from the comments page.

If you want credit, include your name and email (otherwise it's anonymous). You can use HTML but you don't have to.

[Up: imhofaq] [Site map] [Robot Wisdom home page] (Feedback to jorn@robotwisdom.com)

Hosting provided by instinct.org. Content may be copied under Open Web Content License.