Pointless question...

Jules Richardson wrote: ... Typical bit rates for magnetostrictive delay lines used in 1960s-era calculators was on the order of 200 - 600KHz. I did calculations for propagation speed and 'physical bit length' for one of these calcs at one time but don't have the figures at hand. What's the maximum bit rate on optical fiber these days? If one were to utilise the full bandwidth of the fiber to time-slice the fiber into n channels, recirculate the bits sequentially through each of those channels, one could achieve a factor of n reduction in the length of fiber needed. I had to make a variable-length binary delay line a year or two ago for a project to get two 1950s-era Deskfax fax machines communicating (needed storage for one line of the image because the drums were not synchronised, and could start at an arbitrary phase relationship to each other). It was only a couple thousand bits of storage needed but I ended up just using 1 bit-line of an 8*8K static RAM and ignored the other 7 bit-lines. It was much simpler than trying to optimise on a smaller 8*x or 4*x RAM and feed it with shift registers. The variable delay was accomplished with two counters (read and write) started at different times and multiplexed into the RAM address lines.