Star letter

Paul Williamson’s article ( see issue 312, p118) reminded me of my introduction to computing.

He and I are contemporaries, but followed different paths. As a nascent experimental scientist, I began at the Atomic Energy Research Establishment (AERE) in Harwell in 1965 as a scientific assistant, where my introduction to computing and computers was with IBM’s first transistorised “supercomputer”, the IBM 7030 Stretch at the Atomic Weapons Research Establishment (AWRE) in Aldermaston.

Communication with the Stretch was by way of a van, which each day picked up trays of Hollerith cards from Harwell, transported them to Aldermaston and returned the following day with the trays and, hopefully, the printed results. Very quickly, I learned to triple-check the order of the Hollerith control cards to prevent “job failed” printouts returning.

In 1970, I transferred to the newly created Institute of Hydrology at Wallingford, where I first encountered mini-computers (which were still the size of a modern fridge!) in the form of the HP 2116B housed in a caravan ( see image). This contained 8KB of 16-bit-wide memory on individually wired ferrite cores. Built to US military specification, it could be, and very nearly was on several occasions, dropped onto concrete from several feet without damage. It cost $26,000 at that time.

The HP 2116B was used to control, analyse and print data from a set of meteorological sensors on towers above Thetford Forest. Beginning as a mere operator of such a system, , I rapidly involved myself in the production of Fortran coding. LaterLt iin theth 1970s,1970 the th first low-power CMOS microprocessors became available from RCA: the CDP1802 COSMAC, which with a CDP18S601 Microboard computer and a CDP18S643 ADC board allowed our team to develop a battery-powered system for measuring the atmospheric turbulent eddies using novel miniature sonic anemometers and infrared gas analysers.

Without any formal training, I was asked to write the controlling software in assembler, which required a steep learning curve before even being capable of writing code. The requirement was to record, initially at 16Hz, measurements from the sensors, convert, apply calibrations and algorithms, create moving averages and output to book-sized 16KB CMOS memory stores. At 16Hz, this proved impossible – and even at 10Hz, a whole 100ms cycle at the end of each 30-minute measurement period had to be allocated for creating and storing the period averages and variances.

My computing journey continued right through to my retirement in 2007 as an adjunct to my scientific studies, and I increasingly used robust batterypowered data loggers to devise, construct and code increasingly complex measurement systems for use around the world in harsh environments from the Amazon to Sub-Saharan Africa.

Our star letter writer wins a copy of Serif Affinity Designer. Built from the e ground up over a five-year period, every feature, tool, panel and function has been developed with the needs of creative professionals at its core.