Relay computerswhy so few 12

There were other relay calculators. IBM produced two relay calculators for Aberdeen during the war and it produced three more Aberdeen machines two used by the IBM Watson Lab and one other used by the Navy were not as programmable as some of the Bell machines (the most general purpose Model V not finished until after the war was over) but they could do 48 step sequences.

Further afield Konrad Zuse made relay computers both during and after the war. I believe that his company released at least one relay computer in Germany after the war apparently ending with the Z11 after which his company made electronic machines. There were probably other international manufacturers of relay machines and there were certainly one off relay machines built all around the world.

In terms of simpler calculators all IBM mutliplier unit record machines (electronic accounting machines, EAMs) 600-602A used relays to do their multiplication tables as far as I can tell. Perhaps other electomechanical desk calculators used them as well.

I would wonder if a relay computer is any less complicated than a vacuum tube based one. While tubes were more expensive than relays if all the labour and parts to wire them together were the same then a relay and electronic machine might have a price in the same order of magnitude. The electronic machine would be much faster though.

Since IBM switched to electronics for its multpliers in 1948 presumably the costs were not prohibitive and the speed an advantage. Despite the fact that IBM's first multipliers 603 and 604 did not have much memory or sequencing capacity which severly limited the extent to which electronic speed could be taken advantage of.

Note that the first IBM large-scale programmable machine the SSEC used relays as memory and tubes for its arithmetic unit Calculator (a combination of several IBM unit record machines that could follow instructions from a card) used relays as an additional These would be hybrid relay, electronic machines.

As mentioned elsewhere, an organized group of people with desk calculators can do a great deal. Richard Feynman claimed that he could get his organized group of human computers at desk calculators to work as fast as IBM machines (basically ordinary accounting machines) at Los Alamos. Now those machines would be slower than a fully realized relay machine but some of that difference would be made up by possible faster desk calculators. Also, some of that speed is going to be do to the slowness of reading in data mechanically and so hard to make up. Finally, an increase in speed does not necessarily make it cheaper to do the calculation and a small increase in speed will not make an unnattractive calculation attractive.

I am not clear but I thought relay machines could have maintenance troubles including intermittent things like dust and stuff deposting on the contacts. Although vacuum tubes clearly had more.

So relay machines were used but apparently they were already almost obselete thanks to vacuum tubes and the other electronics of the late 40s and early 50s.

Perhaps if someone had tried to develop a computer earlyer in the 30s or 20s relays would have had their time in the sun (then again maybe that would have just lead to making more reliable tubes sooner edging out the relays quickly again). -- Yours Truly, Allan Olley