In 1970, MOS memory chips were just becoming popular, but were still very expensive. Intel had released their first product the previous year, the 3101 RAM chip with 64 bits of storage.[1] For this chip (with enough storage to hold the word “aardvark”) you’d pay $99.50. To avoid these astronomical prices, some computers used the cheaper alternative of shift register memory. Intel’s 1405 shift register provided 512 bits of storage – 8 times as much as their RAM chip – at a significantly lower price. In a shift register memory, the bits go around and around in a circle, with one bit available at each step. The big disadvantage is that you need to wait for the bit you want to come around, which can take half a millisecond.
Like Motorola did :
http://tinymicros.com/wiki/MC14500B
http://tinymicros.com/wiki/MC14500B_In_VHDL
The MC14500 was an interesting beastie. While technically a microprocessor (though you have to implement your own program counter), it is really intended to be used as a PLC for industrial applications.
Shift Registers (in their various forms) are still around today, and still widely used. They got used as RAM replacements because SISO (Serial In, Serial Out) shift registers are very similar to the kind of delay line memory that was used in the early days of computing.
I am not trying to take away from your point, more tring to understand it
I wonder if we could take this concept and translate it to something useful today.
Why not replace this serial shift register memory with a photon propagation delay memory
Let’s see, c=299,792km/s, the moon is 384,403k away and has a corner reflector on it.
http://en.wikipedia.org/wiki/Corner_reflector
This means there are 2.56s of propagation delay. A 10gb/s optical signal would give us 25.6 gigabits of storage in transit to the moon. It’s too bad we’ve got this atmosphere attenuating our signal. Probably take a big telescope, not to mention birds and plans getting in the way.
I love these articles on the history of computing. It’s good to take a little time once in a while to consider where we came from, and other directions things might have taken.
For example, with these shift register memories, even though the latency is a major problem, the article makes an interesting point that the serial nature of the output was fine for things like displays. So in a system that is mostly idle (like calculators or terminals), it may not have been a bad choice (considering the cost difference).
Be careful in counting a technology out. IBM’s Racetrack Memory (http://en.wikipedia.org/wiki/Racetrack_memory) is another variation on delay line/shift register memory and may hold the key to very high memory density with DRAM like access speeds.
The longer you are in the industry, the more you will realize that Nothing Ever Really Dies…