notes.md

Conference FP TS (45 minutes + questions)

Making the compiler work for you

The promise of Typescript

• Leverage the type system to model your domain
• Catch errors at compile time to avoid surprises at runtime.
• Refactor with confidence

Where Typescript falls short

• Working with impure code
• Managing errors
• Manipulating runtime data

The promise of functional programming

• Make the code predictable at all times
  • Emphsizing pure functions
  • Encapsulating side effects
• Reusing code through function composition and categories

How can we benefit from statically typed FP ?

Statically typed FP allows us to write more declarative code, focusing on the what rather than the how.

The type allows us to purify function using algebraic data types rather than exceptions, and model truly total functions. Side effects are clearly identified through the use of IO monads.

This idea, combining functional programming with a strong type system isn't new. It is the fundation of languages like Haskell and ML, and nowadays, it is gaining popularity in the javascript community through compiled languages like ReasonML, Scala.js, PureScript ...

The characteristics of a statically-typed, FP language

• It has support for firt-class functions
• It emphasizes on referentially transparent expressions
• It has ways of handling effectful computations
• It has support for algrebaraic data types (discriminated unions, Either ...)

The characteristics of a pure function

Pure functions should be total

Working with option values

Interoperate with impure code

https://dev.to/gcanti/interoperability-with-non-functional-code-using-fp-ts-432e

Refining the domain

Exceptions break the flow at runtime

Using Either

Should I use Option or Either, the cheatsheet ?

Composing pure functions

Conclusion : how does fp-ts fit in the initial definition ?

What is left to explore

Questions