What is it ? Mechanics and mechanisms
Usage Patterns
Usage sample
Internals
Data Structure
Major Operations
Reference Usages in the Linux Kernel
Experience – what worked and more importantly what didn’t work & gotchas
For tables less than 50 elements or so, binary search will give more performance than hash tables
Collision and how to handle collisions is an important factor in designing around hash tables
There are schemes that generate a perfect hash function, given a set of elements
While there is no clear size, from my experience, 16 bit hashes that give 65536 elements is a good start
You can have global and local hash tables ie smaller hash tables than a big hash tables that contains all the elements
Discussions & Notes
<KS 8 Feb, 07> I have started the template and am slowly filling in the details. Please contribute as you see fit. Cheers </KS>
