| Home > LEO Computers > LEOPEDIA > Other Memoirs, Reminiscences > John Simon Florentin, Computer Operator |
John Simon Florentin, Computer Operator
|
I used to work on Leo III's at Shell Mex's computer centre in Hemel in about 1965-67. I thought I had better write something about this before I disappear for good.
They had two of these machines that were to us exactly the same. All this was on the lowest level of the building where it crossed the bottom end of Marlowes in the centre of Hemel-Hempstead. This building has since been replaced.
These machines were run 24 hours a day Monday to Fridays. There would be two teams of about 4 operators on each shift, one for each machine. The only other persons around on the night shift would be someone in charge of the whole shift, the cook to feed us and a person who was the magnetic tape librarian and the engineers. On the night shift the operators would be fed at about 12.00 at night and the computers would be handed over to the engineers who would do whatever maintenance was necessary. They frequently did not use there allocation of time and it was possible for anyone who wanted to, to run their own programs. I think one machine was handed over to the engineer for an hour then the other for an hour. When the operators were not working or eating one popular pastime was playing cribbage. On the end of the week night shift there would be a very large sort that took most of the shift. But when it finished everyone went home early.
Each machine had eight tape drives (no discs) These would be four on two channels - so the source files would be read in from two drives on one channel and output to two other drives on the other channel. The source file would be on about 10 tapes.
This meant that one person would be working almost by himself juggling these tapes all night. On top of this, every now and then the tapes containing the current partially sorted file would be saved just in case a re-run was needed. During a large sort like this the machine could run up to two other programs
All programs had to be typed onto paper tape and then read in.
I think the words in the memory were 48 bits long. One or two of these were parity bits as operators would get SPF's on the console denoting a store parity failure.
It was said these machines had a floating point option fitted.
The control store was in a box about 1ft cube. Inside was a three dimensional array of fine wires with very small cores at what appeared to be random points.
The paper tape reader was made by Elliot (sic) but the paper tape punch was made by Teletype. The printer was made by Anelex. There was also a card punch and reader. There might have been a reader for reading forms where boxes were selected by drawing a line through them (Frank Land, editor: Lector or Autolector). The whole machine was made using transistors except in the tape drives, TM2's made by Ampex. Thyratrons were used to control the roller used to press the tape against the drive. The density of the bits on the mag tape was such that we had a gadget with a sort of fluid magnetic liquid in it such that individual bits could be seen.
Date : UnknownThis exhibit has a reference ID of CH56459. Please quote this reference ID in any communication with the Centre for Computing History. |
|
