23Z Page 108
23Z HISTORY OF MACHINE SETTING
The original machine methods of setting were naturally the same as the hand statistical method (see Ch. 44). That is to say the Χ runs were of the form i+j/.
The motor runs were all of the form motor = . given ΔD = /. The Ψ. runs were of the form Pi + Pj = ., using a contracted de-chi (see part 4).
The statistics for all this were at first very scanty. Consider for example the surprise expressed in R0, 25 at the failure of a 1+3/ run. (See also R0, 72).
When the Χ2 limitation was introduced it was seen that this was no serious matter and the B.I.'s involving Χ2 were done making use of the limitation by having Χ2 put in the third impulse of the Χ1, 2 tape.
After this it was realised that /'s in ΔD were a good thing to look for, so that Χ4 and Χ5, for example, could be set by the long run 45/123, instead of 4+5, 5+2 etc. This was done by a de-chi of the first three impulses only (R0, 1). Anti-repeats in D were suggested too, as being due to /'s in ΔP at motor crosses. It was only later realised that anti-repeats in P were quite likely to be good (R0, 44, 45).
The idea of making simultaneous use of repeats and anti-repeats occurred first in connection with motor setting (R0, 77).
The run repeats or anti-repeats was an example of the value of being able to use the same electrical impulse more than once. This facility was advocated first in R0, 41. At the same time the 'and/or' machine was advocated. The fact that 'not' can be used as a method of saying 'or' was implied in R0, 23. All this can be regarded as the germ of the idea of the Colossus switchboard. Other suggestions for machine improvements that were suggested in those days but which were adopted only in the sense that they had some slight influence on future methods were
(i) Possible use of Δ2 properties for Χ-setting (August, 1943).
(ii) Decibanning machine (R0, 43)
(iii) Square-summing method for using heterogeneity (R0, 29).
The tendency to think in terms of repeats instead of in terms of the 32 letter count of ΔD is the origin of the use of the use of the symbol r to denote the 'number of places looked at'. r at first was always a number of repeats. This attitude was changed overnight by a single ΔP letter count that was
Back to General Report on Tunny. Contents.