Skip to content

ColinShaw/ocaml-c-map-reduce

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

7 Commits
 
 
 
 
 
 
 
 

Repository files navigation

Peave About Map, Reduce and People Who Talk About 'The Functional Style'

I have a significant grievance with people speaking of functions such as map and reduce in non-functional languages as functional programming. There is nothing special about map and reduce, they are not functional concepts, they are very basic list operations. The whole notion of needing to have a defense for communities that are unwilling to learn what functional programming really is makes me want to pull my hair out.

Reduce is also known historically as fold, which is a more generally used term in functional communities. In some languages it is known as inject, and in imperative languages having enumerations it tends to be termed names more like aggregate. The choice of term tends to depend on the history and lineage of the language. You can generally tell if someone knows what they are talking about relative to functional languages if they can describe the dichotomy of fold and why the two forms matter. For those interested in what is important abot functional programming, there are abundant resources available on the web, or you could look at these slides that I put together for a presentation. For even more information, I would encourage looking into one or more of the resources I put together in this repo.

A good question to pose is why functional implementations admit elegant use of functions like map and fold. There are several answers to this question. First, many functional languages are LISP-like in the sense that list operation are fundamental operations on types, rendering basic language syntax that makes list operations quite expressive. More than that, the basic operations on lists have special properties that facilitate provability of execution. Second, functional languages are applicative and support first-class and higher order functions. That is not something you can fake in a different language, it is deeply part of the design of the language. Stating that an imperative language is functional because there is support for map and fold is simply incorrect. At best these features are an accommodation for the laziness or ineptitude of those who use the language.

I am perfectly happy to embrace that there are positive features in any language to using methods entitled map, reduce and others on iterables, enumerables, collections or whatever one wants to call indexable multiples since it significantly reduces error-prone human behaviors when manually indexing things. This is a huge improvement, as it takes the error opportunity out of rote code. However, you can't say it is functional, it is not, it is just not true. Like I said before, functional means list constructs coupled with first-class and higher-order functions. You can mimic the list construct, but you are creating a whole new language to support the first-class and higher-order functions.

Herein you will find an implementation of map and fold implemented in both OCaml and C. Yes, C, because doing so helps highlight that these are not functional since C is about as imperative as you can get with a structured language.

You can run the OCaml example by:

ocaml peave.ml

You can build and run the C example by:

gcc peave.c; ./a.out

Both programs will simply return unless there is an assertion error.

About

Example of grievance with people claiming things are functional in their favorite imperative language

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published