The transition to the computer, 1945 to 1951 (continued)

The ideas for, and development of, the Mk 1 were separate from the mainstream of progress overseas some 10,000 miles away. It was known from the ASCC experience that use of paper tape for providing instructions for sequential execution at run time was too slow, that data and instructions had to be available for execution and operation at comparable speeds. This required a means of holding and readily accessing both data and instructions. Also, decoding and encoding processes and a scheme for specifying and sequencing execution of sets of instructions, i.e. a programming technique, had to be devised. Pearcey was aware, however, of the neurological work of W. S. McCulloch and W. Pitts (1943) and of T. Gold's work at Haslemere on the use of acoustic delay lines for radar signal enhancement and the possibility of this as a storage medium.

The Mk 1 was a 'programmer's machine, structured for engineering simplicity and economy and flexibility since the main objective was the development of programming techniques leading later to breadth of application. The machine was entirely serial in operation, with 20 bits (binary digits) per instruction/number (word) and performed about 500, later 1000, operations per second using a mercury acoustic delay store of up to 1024 words. Extensions included an additional 4096 words magnetic drum storage. Input and output, initially via punched cards was changed to 12 hole and standard 5 hole paper tape. The machine was first demonstrated publicly in August 1951 and was used for programming studies, computations for radioastronomy and cloud physics for RP researches and computations for various CSIRO Divisions, engineering projects, various university and government departments and agencies in Sydney until mid-1956 when it was transferred to the University of Melbourne. Late in 1948 Myers visited some of the UK computer projects at CSIR'S request and, in 1949, advised that work on development of computing machinery should be confined to components only -at least until a cheap and reliable directly accessible storage medium became available. This was to affect what followed, particularly in respect of the Mk 1.

This period ended with the first Conference on Automatic Computing Machines, held in August 1951 at the Department of Electrical Engineering and sponsored by that Department and CSIRO. Thirteen papers were given to 160 participants. The conference was accompanied by an exhibition demonstrating use in scientific computation of various commercially available calculating and accounting machines. The Mk 1 computer and Hollerith (BTM) punched card system at RP and the electromechanical differential analyser at the Department of Electrical Engineering were included.

Australian Academy of Technological Sciences and Engineering