Martlet

From ESIWiki

Martlet is a programming model developed by Daniel Goodman as part of his doctorate under the supervision of Andrew Martin at the Oxford University Computing Laboratory. This work is inspired by the inductive constructs of Functional programming and the data analysis requirements of ClimatePrediction.net. Martlet uses these inductive constructs to allow users to write functions that can adjust automatically to the distribution of data in a dynamic environment, so removing the need for either user intervention or large data transfers when such distributions change.

Martlet is different to other workflow language because as well as changing the scheduling of tasks to suit the environment it also changes the abstract syntax tree of the function, while maintaining ssemantic correctness. This is done through the addition of two classes of data structure which collectively abstract this partitioning from the user.