Anonymous user
Register Machine Model of Computation: Difference between revisions
Register Machine Model of Computation (view source)
Revision as of 22:24, 25 November 2017
, 6 years ago→Computational Models and the History of Computing
| [unchecked revision] | [unchecked revision] |
m (appearance) |
|||
Line 12:
The next significant model to arise was the ''lambda calculus'', created by Alonzo Church in 1936. This was a simpler model of functions than the existing type of recursive function theory, but was demonstrated to have the same computational power as RFT. That same year, Alan Turing developed yet another model, the ''Turing Machine'', which he used to demonstrate that a well-known unsolved goal in mathematics — finding a method for proving that any arbitrary computation process would go to completion — was impossible, by showing that the testing process itself would necessarily be undecidable in the general case. In doing so, he showed that it was possible to design a variant of the Turing Machine that could simulate any other possible Turing machine; he postulated that this ''Universal Turing Machine'' could perform any decidably mechanizable computation. -
These models were 'mechanical' only in the sense of being built on a set of rules; they were not very useful for developing a physical computing device. Of them all, the UTM was the closest, as it was described in terms of an imaginary mechanism consisting of a reading and writing head and an infinitely long paper tape. While it was not a practical design, it demonstrated that the mechanical calculators of the time had a solid theoretical basis, and more importantly, re-introduced the concept (originally proposed by Charles Babbage) of a ''programmable'' computation device.
Turing would go on to do significant work in the development of one of the first such machines, the ULTRA project's Colossus decryption machine originally conceived by Tom Flowers in early 1943. During this period, from the late 1930s to the early 1950s, several groups of designers seem to have developed similar ideas independently of each other: Konrad Zuse in Germany, John Atanasoff and Clifford Berry at Iowa State University, Presper Eckert and John Mauchly at Princeton, and Howard Aiken at Harvard, all rediscovered the principle of stored program computers between 1941 and 1943, a concept later codified in a white paper by a team led by John Von Neumann. Because of this, designs in which the program and data are stored in a shared memory became known as the ''Von Neumann architecture'', while those which have independent memories for program and data, such as the Zuse Z4 and the Harvard Mark 1, became known as the ''Harvard Architecture''.
| |||