A dual calculus for unconstrained strategies

Add to your list(s) Download to your calendar using vCal

• Alexis Goyet, PPS
• Thursday 25 April 2013, 16:00-17:00
• CS 217.

If you have a question about this talk, please contact Dan Ghica.

The aim of the lambda lambda-bar calculus is to represent as faithfully as possible the structure of non-deterministic, non-innocent, non-visible strategies in game semantics. Non-innocent strategies have been shown to model functional languages extended with references. The typical approach is to model the purely functional fragment with innocent strategies, to which a “memory-cell” strategy is added. Instead, our starting point is a simple concrete syntax of finite strategies, which are inherently non-innocent.

The resulting language is a dualization of the lambda calculus. A new binder, the lambda-bar, names arguments which have been passed to the environment (whereas the lambda names arguments which have been received). This makes explicit the act of answering a function call, and allows this answer to change over time, granting the power of effects.

This talk is part of the Lab Lunch series.

This talk is included in these lists:

• CS 217
• Computer Science Departmental Series
• Computer Science Distinguished Seminars
• Lab Lunch
• Theoretical computer science seminar

Note that ex-directory lists are not shown.