TypeScript and ECMAScript have some of the most beautiful constructs and syntaxes to do parallel, non-blocking programmingâ€"but unfortunately, only for standard APIs like fetch in the browser or file IO in Node. When implementing anything CPU-bound in TypeScript, offloading it into a different, parallel execution context and exposing an equally type-safe and flexible API for it is non-trivial.
TypeScript and ECMAScript have some of the most beautiful constructs and syntaxes to do parallel, non-blocking programming—but unfortunately, only for standard APIs like fetch in the browser or file IO in Node. When implementing anything CPU-bound in TypeScript, offloading it into a different, parallel execution context and exposing an equally type-safe and flexible API for it is non-trivial.
At the 2019 TSConf pre-meetup in Seattle, I presented the different options for implementing parallel APIs in userland, starting from any-typed message passing, to utilizing advanced TypeScript features such as mapped types, conditional types and function parameter tuple types to support remote procedure calls, fully-flexible property access, and live references to mutable objects.
After watching this talk, you'll have learned how to implement truly-parallel CPU-bound APIs in userland, and how to expose them to other threads, Node processes or even over the network in a fully type-safe manner, all with zero boilerplate code.
With Sourcegraph, the code understanding platform for enterprise.
Schedule a demo