Algorithms and stuff
Many thanks to “Paul”:http://matrix.netsoc.tcd.ie/~paul/ for pointing out skip lists as an alternative to b-tree implementations like AVL trees. Much less complexity in programming, and similar performance in the normal case to highly tuned AVL algorithms. Skip lists were developed relatively recently (a paper was presented to the ACM in 1990) by “William Pugh”:http://www.cs.umd.edu/~pugh/. I’ll have to actually implement a version of this in a real product (again, by replacing a linked list structure) and see what happens 
The next frontier after that is text indexing. “Modern Information Retrieval”:http://www.amazon.com/exec/obidos/tg/detail/-/020139829X/qid=1082639150/sr=8-1/ref=sr_8_xs_ap_i1_xgl14/103-9422780-4738264?v=glance&s=books&n=507846 seems like just the ticket for a good introduction to the area. I’ve also been reading some articles about inversions and text signatures in “Citeseer”:http://citeseer.ist.psu.edu/ (which is an amazingly useful resource for this kind of research).
Maybe some day I’ll be given some time in my day job to work at this For the moment, it’s mostly limited to some investigative reading during lunchtime and at weekends.