o_ComputeHashFromTo

Compute the hash value from the object's data. The result should be an int. To keep the rehashing cost down for concatenated tuples, we use a non-commutative hash function. If the tuple has elements with hash values $h_1{h}_n$ , we use the formula $h_1{a}^1+h_2{a}^2++h_n{a}^n/$. This can be rewritten as $\underset{i=1}{\overset{n}{}}{h}_i{a}^i$ ). The constant a is chosen as a primitive element of the group $\left({}_{2^{32}},\right)$, specifically 1,664,525, as taken from Knuth, The Art of Computer Programming, Vol. 2, 2^nd ed., page 102, row 26. See also pages 19, 20, theorems B and C. The period of this cycle is 2³⁰.

To append an (n+1)^st element to a tuple, one can compute the new hash by adding $h_{n1}{a}^{n1}$ to the previous hash. Similarly, concatenating two tuples of length x and y is a simple matter of multiplying the right tuple's hash by $a^x$ and adding it to the left tuple's hash.

The element hash values are exclusive-ored with a randomly chosen constant before being used, to help prevent similar nested tuples from producing equal hashes.

Parameters