IBMWatson autobiographythoughts on 779
IBMWatson autobiographythoughts on 780
btw ... has anybody actually seen a hardcopy? they had a photographer come out for a photoshoot ... but pictures don't show up in the online version. note that it wasn't...
How much RAM is 64K 36bit words of Core Memory 785
Paul' "I get confused with the old core memory descriptions of memory being WORDS rather...
You have a point, still the number of employees from 1920 to 1929 is double (2,731 to 5,999) and actual net earnings more than tripled ($2 mil to $7 mil), although total revenues were only up a bit (14 mil versus $18 mil.) (numbers from the IBM Archives site the year and the numbers for the year are put up in a sidebar). I think that is pretty good growth for a company. Enough to suggest that there was some real solid demand for the new accounting practices even before the complexities of new payroll taxes were created.
The SSEC's arithmetic unit was actually fully electronic (using the same basic vacuum tube circuit designs as the 604). The relays were used for memory. In a sense the 604 was a step back form the SSEC because it had little fast memory (you had to punch cards) and I get the impression you had to plug wire program to a great degree. The CPC units included extra memory in the form of relays. I do not think it is the inherent workability of the technology that is at issue. Electronic machines are more complicated than mechanical machines in many senses.
I think the Mark I and the pure relay machines (Bell Labs Mark V, IBM pluggable sequence relay calculator-Aberdeen machines) were pretty useful machines, but perhaps not as economical as they might have been and certainly left in the dust by electronic machines.
However, technology is clearly an issue. L. J. Comrie found in the 1930s that for most applications well organized use of human computers with desk calculators would be cheaper than the IBM type punched card operations. WJ Eckert used EAMs extensively in his work in the 30s but since he was getting them gratis from IBM it may be that he would have been better off getting an equivalent money grant and running a room full of people with desk machines. I still suspect that some of the work was good return on costs (might have been better if the applications had been chosen with a more pecunary interest). Comrie's own use of Hollerith machines to do the lunar ephemeris in 1932 was by his calculation a significant savings, the most telling thing to me about that was that he continued the calculations until 2000 because it was so little extra effort.
Arguably it comes downt to Grosch's Law (cost of computation goes down in proportion to the square root of speed), you have to increase the speed significantly in order to really be doing your calculations any cheaper (since Grosch based his law on figures including human computers, punched card machines, the Mark I and SSEC) . Until electronics comes along automatic machines are not significantly faster than humans using desk calculators so the cost Debt Reduction is minimal if it exists at all.
My own sense is that if we were talking about Babbage, he would have been able to build his machines but they would have been white elephants. You could just hire rooms full of people with log tables or what have you and organized them well and achieved the same results for less money. Whereas by the 20s or 30s the potential to get some real use out of automatic calculation was there both in data processing and in straight calculations. So I think there might have been room for more gains than people like WJ Eckert were able to ring out of the machines of the day. Still the potential benefits were not any where near as great as those seen after the developments of electronics. Also, the demand in terms of explicit engineering and research applications clearly increases with WWII and attendant technological developments. -- Yours Truly, Allan Olley
IBMWatson autobiographythoughts on 782
ibm would frequently get the science and engineering of manufactoring production extremely optimized (lots of studies of the manufactoring process, yields, quality, optimizations, etc) ... so the...