[Pl-seminar] 2/24 Seminar: Ben Sherman, Overlapping pattern matching for programming with continuous functions

[Daniel Patterson]

NUPRL Seminar presents

Ben Sherman
MIT
Host: Gabriel Scherer

12:00pm
Friday, February 24, 2017
Room 366 WVH (http://www.ccs.neu.edu/home/wand/directions.html)

Overlapping pattern matching for programming with continuous functions

Abstract:

Topology, when viewed from an unusual perspective (formal topology),
describes how to compute with certain objects that are beyond the reach of
induction, such as real numbers, probability distributions, streams, and
function spaces. Accordingly, one gets a programming language whose types
are spaces and whose functions are continuous maps.

Does this language have pattern matching? In functional programming,
pattern matching allows definition of a function by partitioning the input
and defining the result in each case. We generalize to programming with
spaces, where patterns need not represent decidable predicates and also may
overlap, potentially allowing nondeterministic behavior in overlapping
regions. These overlapping pattern matches are useful for writing a wide
array of computer programs on spaces, such as programs that make
approximate computations or decisions based on continuous values or that
manipulate "partial" datatypes.

This talk will introduce topology from a computational perspective, and
explore some programs that can be built with this framework using
overlapping pattern matching.

Bio:

Ben Sherman is a second-year PhD student at MIT, advised by Adam Chlipala
and Michael Carbin.
-------------- next part --------------
HTML attachment scrubbed and removed