Introducing Zod Codecs

I’ve been working for a while (unpublished yet) on the notion of (sometimes bidirectional) projections, and how they can help with data mapping and reshaping. It’s good to see others taking steps in that direction. So much code in backends and frontends is spent in reshaping/assembling/deassembling and yet the tooling is quite primitive still.

This reminds me a lot of database lenses from ink & switch: https://martin.kleppmann.com/2022/04/29/hytradboi.html Sadly they never published an implementation. Can anybody explain the differences or if they build on similar ideas?