While I appreciate the Douglas Adams tribute, I wanted to raise the possibility of (1) changing the default murmurhash seed value to 0 (default in many bindings) and/or (2) exposing it as a command line flag.

The former would be useful for compatibility (both cross lang/lib, and with existing k-mer hashing code that uses murmurhash3). The latter may also be useful in facilitating easier simulation of many pseudorandom minmer subsets from the same file, and for interop with code that doesn't use a seed of 42.

Obviously (1) is a highly breaking change and therefor good to discuss and pin down prior to a preprint coming out. Thoughts on the marbl side?