Heap Compression and Binary I/O in Haskell

Two new facilities for Haskell are described: compression of data values in memory, and a new scheme for binary I/O. These facilities, although they can be used individually, can also be combined because they use the same binary representations for values. Heap compression in memory is valuable because it enables programs to run on smaller machines, or conversely allows programs to store more data in the same amount of memory. Binary I/O is valuable because it makes the file storage and retrieval of heap data structures smooth and painless. The combination of heap compression and binary I/O allows data transfer to be both fast and space-efficient. All the facilities described have been implemented in a variant of Röjemo’s nhc compiler. Example applications are demonstrated, with performance results for space and speed.