Programming with Managed Time

TL;DR: Glitch is introduced as a form of managed time that replays code for an appearance of simultaneous memory updates, avoiding the need for manual order and restricting the imperative model while retaining the basic concepts of memory access and control flow.

Abstract: Most languages expose the computer's ability to globally read and write memory at any time. Programmers must then choreograph control flow so all reads and writes occur in correct relative orders, which can be difficult particularly when dealing with initialization, reactivity, and concurrency. Just as many languages now manage memory to unburden us from properly freeing memory, they should also manage time to automatically order memory accesses for us in the interests of comprehensibility, correctness, and simplicity. Time management is a general language feature with a large design space that is largely unexplored; we offer this perspective to relate prior work and guide future research.We introduce Glitch as a form of managed time that replays code for an appearance of simultaneous memory updates, avoiding the need for manual order. The key to such replay reaching consistent program states is an ability to reorder and rollback updates as needed, restricting the imperative model while retaining the basic concepts of memory access and control flow. This approach can also handle code to enable live programming that incrementally revises program executions in an IDE under arbitrary code changes.