CSKIPLIST

(c) Robert Winkler 2011
=======


This is a simple implementation of a skiplist.  I originally
planned to make a more complete version like CVector but
trying to design a skiplist implementation that takes advantage
of all the variations and optimizations of skiplists and presents
a clean, easy to use API seems virtually impossible.  Since I design
it with how I would use it in mind, I thought it'd be better to just
make these 2 simple versions and if I ever need to, I can add to
these and modify them with the extensions/optimizations.

Some of the things you can change/select/implement for skiplists are

P value 
K value (P and k can be used to optimize for size vs speed)
Max level (= log_1/P(N) where N is max number of elements)
Whether to allow duplicates (and how to deal with them either way)
Whether to allow arbitrary key types and allow user to specify compare function
Optimize for locality of reference searches using search fingers
Optimize for expensive comparisons
Modify to work like a vector, accessing elements by index
Merge, Concatenate, Split algorithms

etc.

You can read all about skiplists on wikipidia and read some
of William Pugh's papers and see example code for a basic implementation
here:

ftp://ftp.cs.umd.edu/pub/skipLists/

I've run it under valgrind so there are no memory leaks/problems.  My tests
all pass but they could probably be more thorough.