I have not progressed much lately on this front but I just finished writing a new article, due for publication next month probably.

In there, I develop a little new theory derived from Bob Jenkin's funnel principle, but applied to the dataflow graph of the checksums.

a) is obviously the very flawed checksum algorithm.

b) is the Fletcher algo, not quite there.

c) is PEAC which differs from Fletcher by a very little rewiring of a feedback operand...

How come the Fibonacci structure is not the de facto standard ?