Map and Bind and Apply, a summary

Series summary

Well, this series turned out to be longer than I originally planned. Thanks for making it to the end!

I hope that this discussion has been helpful in understanding the various function transformations like map and bind, and given you some useful techniques for dealing with world-crossing functions -- maybe even demystified the m-word a bit!

If you want to start using these kinds of functions in your own code, I hope that you can see how easy they are to write, but you should also consider using one of the excellent F# utility libraries that contain these and much more:

• ExtCore (source, NuGet). ExtCore provides extensions to the F# core library (FSharp.Core) and aims to help you build industrial-strength F# applications. These extensions include additional functions for modules such as Array, List, Set, and Map; immutable IntSet, IntMap, LazyList, and Queue collections; a variety of computation expressions (workflows); and "workflow collections" -- collections modules which have been adapted to work seamlessly from within workflows.

• FSharpx.Extras (home page). FSharpx.Extras is part of the FSharpx series of libraries. It implements several standard monads (State, Reader, Writer, Either, Continuation, Distribution), validation with applicative functors, general functions like flip, and some asynchronous programming utilities, and functions to make C# - F# interop easier.

For example, the monadic traverse List.traverseResultM that I implemented in this post is already available in ExtCore here.

And if you liked this series, I have posts explaining the State monad in my series on "Dr Frankenfunctor and the Monadster" and the Either monad in my talk "Railway Oriented Programming".

As I said at the very beginning, writing this up has been a learning process for me too. I am not an expert, so if I have made any errors please do let me know.

Thanks!

Series contents

Here's a list of shortcuts to the various functions mentioned in this series:

• Part 1: Lifting to the elevated world
  • The map function
  • The return function
  • The apply function
  • The liftN family of functions
  • The zip function and ZipList world
• Part 2: How to compose world-crossing functions
  • The bind function
  • List is not a monad. Option is not a monad.
• Part 3: Using the core functions in practice
  • Independent and dependent data
  • Example: Validation using applicative style and monadic style
  • Lifting to a consistent world
  • Kleisli world
• Part 4: Mixing lists and elevated values
  • Mixing lists and elevated values
  • The traverse/MapM function
  • The sequence function
  • "Sequence" as a recipe for ad-hoc implementations
  • Readability vs. performance
  • Dude, where's my filter?
• Part 5: A real-world example that uses all the techniques
  • Example: Downloading and processing a list of websites
  • Treating two worlds as one
• Part 6: Designing your own elevated world
  • Designing your own elevated world
  • Filtering out failures
  • The Reader monad
• Part 7: Summary
  • List of operators mentioned
  • Further reading

Appendix: List of operators mentioned

Unlike OO languages, functional programming languages are known for their strange operators, so I thought it would be helpful to document the ones that have been used in this series, with links back to the relevant discussion.

Appendix: Further reading

Alternative tutorials:

• You Could Have Invented Monads! (And Maybe You Already Have).
• Functors, Applicatives and Monads in pictures.
• Kleisli composition ? la Up-Goer Five. I think this one is funny.
• Eric Lippert's series on monads in C#.

For the academically minded:

• Monads for Functional Programming (PDF), by Philip Wadler. One of the first monad papers.
• Applicative Programming with Effects (PDF), by Conor McBride and Ross Paterson.
• The Essence of the Iterator Pattern (PDF), by Jeremy Gibbons and Bruno Oliveira.

F# examples:

• F# ExtCore and FSharpx.Extras have lots of useful code.
• FSharpx.Async has map, apply, liftN (called "Parallel"), bind, and other useful extensions for Async.
• Applicatives are very well suited for parsing, as explained in these posts:
  • Parsing with applicative functors in F#.
  • Dive into parser combinators: parsing search queries with F# and FParsec in Kiln.