cleff VS hpqtypes-effectful

I ended up using a different approach to guiding inference in my project, but I still thought the IP way was cool

Consider 2 Haskell projects: Cleff and Effectful. They both are extensible effect systems, built around ReaderT design pattern. At their core, they use fused ReaderT + IO monad, with extensible environment. This architecture is very similar to ZIO's. All 3 even share the same limitation - inability to express nondeterminism (not very relevant fact, but I just want to emphasize the similarity).

cleff and effectful are very similar, to the point that I offered to join forces (for the record, majority of the conversation refers to things that are no longer true for both libraries as we inspired each other to fix most quirks :). The main difference is the internal environment - in effectful it's mutable, while in cleff it's immutable. All differences pretty much grow from this. The main ones:

I'm pleased to announce cleff, a new extensible effects library for Haskell: https://github.com/re-xyr/cleff.

That depends on the API of the library in question. For example, there's hpqtypes-effectful that gives you a DB effect and functions that make use it can perform queries that will by default run in a single database transaction.

For a more complicated real-world scenario you can have a look at https://github.com/haskell-effectful/hpqtypes-effectful.