There’s a style of programming that I have come to love that allows you to think like pipes in a prompt. I figured it out after trying to understand the Schwartzian Transform, which came about from Randal Schwartz answering questions on UseNet in the 1990s. The problem was that someone wanted to sort this data by last name.

adjn:Joshua Ng
adktk:KaLap Timothy Kwong
admg:Mahalingam Gobieramanan
admln:Martha L. Nangalama

And Schwartz suggested this.

#!/usr/bin/perl

require 5; # new features, new bugs!
print
    map { $_->[0] }
    sort { $a->[1] cmp $b->[1] }
    map { [$_, /(\S+)$/] }
    <>;

I have explained this style before, and the key is to read it from the bottom up.

  • <> means we’re getting one entry at a time from STDIN
  • map { [$_, /(\S+)$/] } means we’re turning the string adjn:Joshua Ng into an anonymous array ["adjn:Joshua Ng","Ng"]
  • sort { $a->[1] cmp $b->[1] } sorts the whole thing on the 2nd value in each array
  • map { $_->[0] } undoes the split into anonymous arrays, making an array of strings again
  • print prints them

You could have made this into a series of painstaking steps my @array = <>; my @array2 ; for my $e ( @array ) {...}, to introduce more modern Perlishness into this, but this is clean, by which I mean there are not a series of named variables sitting around.

So, I’ve been writing a lot of JavaScript these days, and have been learning some about arrow functions and other JS hackishness.

"use strict";
console.log(
  function() {
    /*
adjn:Joshua Ng
adktk:KaLap Timothy Kwong
admg:Mahalingam Gobieramanan
admln:Martha L. Nangalama
  */
  }
    .toString()
    .split(/\/\*/)[1]
    .split(/\*\//)[0]
    .split(/\n/)
    .filter(x => String(x).match(/\w/))
    .map(x => [x.split(/\s/).pop(), x])
    .sort((a, b) => (a[0] > b[0] ? 1 : 0 - 1))
    .map(x => x[1])
    .join("\n")
);

Which outputs:

admg:Mahalingam Gobieramanan
adktk:KaLap Timothy Kwong
admln:Martha L. Nangalama
adjn:Joshua Ng

This isn’t bottom-to-top, so

  • we make an anonymous function that is entirely a comment block holding our data
  • which we cast as a String with .toString()
  • use .split() to get the data part of the data by splitting on the open and close comment
  • make into an array by .split() on the newline character
  • .filter() to ensure each entry has a word character
  • use .map() to make an anonymous array, this time with the key coming first, by splitting on string characters and .pop()-ing off the last element as the last name
  • doing the sort, using a ternary operator within .sort() to save space
  • using .map() to get back to the original string we want sorted
  • and finally, making it into a string join .join() with newlines.

Funny that there was no mention of JavaScript in that Wikipedia article.

There are other things we can use to add to this style. I found out earlier today that you can kinda get to Perl’s range operator, like 0 .. 9 , by doing creating an anonymous array, declaring the size and getting a list of the indexes by abusing the object nature, like:

let ranged = [...Array(10).keys()]
# [ 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 ]

And, of course, you can map and filter these things into exactly the shape you want.

If you have any questions or comments, I would be glad to hear it. Ask me on Twitter or make an issue on my blog repo.