I built fp-pack, a small TypeScript functional utility library focused on pipe-first composition.
The goal is to keep pipelines simple and readable, while still supporting early exits and side effects — without introducing monads like Option or Either.
Most code uses plain pipe/pipeAsync. For the few cases that need early termination, fp-pack provides a SideEffect-based pipeline that short-circuits safely.
I also wrote an “AI agent skills” document to help LLMs generate consistent fp-pack-style code.
Feedback, criticism, or questions are very welcome.
The main difference is that SideEffect isn’t a compositional context — there’s no bind/flatMap, and composition intentionally stops once it appears. It’s meant as an explicit early-exit signal in pipe-first code, not a general computation container.
The point of monads is that they solve this exact category of problem in the simplest possible way.
It tries to borrow function composition and declarative structure without requiring familiarity with full FP abstractions like monads or effect systems.
Once you commit to a real monad, you need map/flatMap, lifting, unwrapping, and rules about staying inside the context across the whole pipeline. At that point, the pipe abstraction stops being the primary mental model — the monad does.
SideEffect deliberately avoids that. It keeps the pipe interface intact and only adds a single, explicit signal: “stop here”. That’s why it’s less powerful than a monad, but also much simpler to integrate into existing pipe-based code.
It’s intentionally not a monad, and I’m curious how others feel about this trade-off compared to Option/Either in real-world TypeScript codebases.
Early termination is the most common use case, but it’s not the only thing SideEffect represents. The name is intentionally a bit broader — it’s meant to model “effects where normal composition should stop”.
In practice, that includes things like validation failures, logging or notifications at pipeline boundaries, and error reporting or metrics. That said, the scope is deliberately conservative.
SideEffect isn’t meant to be a general-purpose effect system. If it were, it would quickly turn into something very close to a monad or effect framework, which I’m intentionally avoiding.
https://github.com/superlucky84/fp-pack?tab=readme-ov-file#s...
Still, this approach is very useful for most business logic, too bad most programming languages don't provide a nice syntax for this.
The SideEffect pattern is intentionally explicit, so without fp-pack context it can look unfamiliar at first. The trade-off is making early exits visible in the code, rather than hiding them in conditionals or exceptions.
In practice, most code stays in plain pipe/pipeAsync. SideEffect is meant for a small number of boundary cases only.
And I agree — better language-level syntax for this kind of pattern would make it much easier to adopt.
This happens because `runPipeResult` defaults its generic parameter to `R = any`. When type safety matters, the intended solution is to use `pipeSideEffectStrict`, which preserves all possible SideEffect result types as a precise union throughout the pipeline.
The default version prioritizes ergonomics and simplicity, while the strict version prioritizes type safety.
Also, fp-pack is still in an early stage, so the usability and API choices haven’t been fully validated yet. That’s why feedback like this is especially helpful in shaping the direction of the library.
There’s often a gap between what feels conceptually clean and what teams are actually willing to carry cognitively. Rx in particular tends to exceed that budget pretty quickly.
That’s why fp-pack is intentionally narrow — it’s closer to making a few control-flow cases explicit in pipe-first code than introducing a broad new abstraction.
Please tell me you didn't just add SideEffects to a language full of side-effects.
The intent isn’t to add more side effects to an already side-effectful language. It’s closer to the opposite: trying not to handle side effects all over the place, but to surface them as part of a single, explicit flow.
This is less about adding something like Option to a language without nulls, and more about making control-flow boundaries visible in a multi-paradigm language where effects already exist.
It’s not an attempt to pretend the language is pure, just a small step toward more declarative discipline.