🚀 go-pugleaf

I've also seen some old time books that described memory elements
that stored 1/0 on electrodes by forming nonconductive gas bubbles,
then breaking the bubbles with an applied pulse.  I beleive the
idea of operation is similar to a so-called coherer dectector that
was used in the early days of wireless.  In the latter, a direct
current forms the bubble and the impinging radio signal breaks the
bubble.  When the bubble breaks, a click is heard in headphones
attached across the device.

Re: delay lines.  I worked taking care of the DP equpment in a bank
proof and transit department in the late 1970s.  They were still
using CA 1962 IBM 1260 check inscribing mahcines.  The machines
were all solid state, using IBM familar square silver "SLT"
integrated circuit modules.  ...that is, save the memory.  The
memory was a big clock spring (looks essentially the same as the
sping in my wall clock) sealed in an aluminum box.   There were
transducers at each end of the spring and amplifiers and squarers
to reshape and time the pulses.  Occasionally one of the workers
would trip and bump into a machine and dump the totals, very much
annoying the worker running the mahcine.  The spring memory would
hold about 4K of data.  The system to format the memory on power-up
was pretty interesting.

Before that I bought an early 1960s Olivetti Programma 101 desk
calculator.  It was about the first all electronic solid state
programmable calculators around.  It had about 400 discrete
transistors.  The memory was a few K (forget exactly how much now)
glass rod acoustic delay line.  There was off line memory that used
tabulator cards that had a magnetic strip.  I couldn't find the tab
cards, so I made my own out of mylar with 1/2 inch mag tape glued
on for the stripe; it more or less worked.

-- 
Bill Mayhew      NEOUCOM Computer Services Department
Rootstown, OH  44272-9995  USA    phone: 216-325-2511
wtm@uhura.neoucom.edu   (140.220.1.1)