A Procedure To Harden Salting

Yesterday I piloted the idea of using vector matrices to defeat salting. I noted that the salt need not be in a small range; this might still fall to investigations of multiple instances. Whatever variety of possibilities, the step from one to the next in a series will be constant.

If we defeat salting, we might defeat the effective benefit of hashing altogether. I advance a possible solution from the memory of an embarrassing mistake.

If my salt has become compromised, I can seek a remedy by introducing variations between steps. This is not what we expect from a deterministic hash, so what am I suggesting? 

In the algorithm, the stepping is by one pass of the hash algorithm. Although it is inefficient, I can vary the stepping by taking two hexadecimal numerals from a fixed offset within my hash, converting them to an integer, and choosing to step by a number of hash passes equal to that integer.

Because the hash is deterministic, the resulting list would always be the same, but the vector algebra problem would be multiplied.

Pseudorandom stepping solves the problem; we can experiment with other sources of pseudorandomness.

A unavoidable contraindication is that ultimately, our product key checking slows down.