Functional Self-AdjustingProgramming
Stc
Date: 2009-04-02
Time: 11:45
Room: BBL room 471
Speaker: Xavier Pardonnet
Title: Functional Self-Adjusting Programming
Abstract
The idea of self-adjusting programs comes from the observation
that many applications need to interact with data that changes slightly or
slowly over time. With standard computation techniques, a program has to
recompute the result from scratch, for each and every modification to the
initial data.
Wouldn't it be nice if the program could re-use part of its previous
computation to modify only the parts of the result that has to change?
This is what self-adjusting programming tries to achieve.
In this talk I'm going to present the work of Umut Acar
(
http://ttic.uchicago.edu/~umut/), who has come up with a nice, elegant
and efficient solution to bring this mechanism in the functional
programming world.
