Linear Hashing Work-Around Potential

It is true that we may be talking about patents that may have expired already. And even '120 will have expired no later than 2018 (probably sooner).

However, '120 does apply to linear hashing today if that linear hashing implementation deals with automatically-expiring records and it removes some or all of those on a list of same-hash records while carrying out an operation of the "record search means utiliizing a search key to access a linked list of records having the same hash address."

Considering how fruitful linear hashing has been seen to be, especially for distributed databases, I would still want to make sure I understood about applicable patents before distributing code using the technique (and especially if automatically-expired records are handled opportunistically as part of the split and condense process).

Linear Hashing is used in Berkeley DB in both the free and commercial versions. I actually have no need for a commercial one myself and would use an open-source one if I distributed an application that needed a DB.

I am surprised that Pick handles oveflow buckets the way you summarized it. I see no reason to ever have collisions with the expanding buckets of the linear hash directory (but I am looking only at the in-memory linked-list form at the moment).