As it happens, he’d already done some work on coding theory—in the area of biology. The digital nature of DNA had been discovered by Jim Watson and Francis Crick in 1953, but it wasn’t yet clear just how sequences of the four possible base pairs encoded the 20 amino acids. In 1956, Max Delbrück—Jim Watson’s former postdoc advisor at Caltech—asked around at JPL if anyone could figure it out. Sol and two colleagues analyzed an idea of Francis Crick’s and came up with “comma-free codes” in which overlapping triples of base pairs could encode amino acids. The analysis showed that exactly 20 amino acids could be encoded this way. It seemed like an amazing explanation of what was seen—but unfortunately it isn’t how biology actually works (biology uses a more straightforward encoding, where some of the 64 possible triples just don’t represent anything).

Despite all his contributions to the infrastructure of the computational world, Sol himself basically never seriously used computers. He took particular pride in his own mental calculation capabilities. And he didn’t really use email until he was in his seventies, and never used a computer at home—though, yes, he did have a cellphone.

in June 1955 he wrote his final report, “Sequences with Randomness Properties”—which would basically become the foundational document of the theory of shift register sequences.