I just found a great example of the variability of the need of creating maintainable code.

Somewhere around 2006, I created a script that posts to the Livejournal community daily_tao a new chapter of the Tao Te Ching every day. Today in 2016, I had a need to look for the script, the second time in maybe the last eight years that I had to look at it.

I temporarily wasn't sure of where that script was. I remembered it was running off of a cronjob, so… Some machine with a *nix-like OS. Then I remembered and found it. I'm going to offer two versions of the code to you: the one that is in the cronjob, and .. A rewrite that I never deleted (and now of course, I have no idea why I never did).

The original:

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
  #!/bin/bash

  cd /home/user/www/tao
  TAO_COUNTER=`cat .taocounter`

  if (( $TAO_COUNTER > 80 ))
     then TAO_COUNTER=0
  fi


  awk '/^%$/{nr++} nr==sec,/^%$/{if(!/^%$/)print}' sec=$TAO_COUNTER tao.txt > .body
  echo '-' >> .body
  echo '-' >> .body
  awk '/^%$/{nr++} nr==sec,/^%$/{if(!/^%$/)print}' sec=$TAO_COUNTER beatricetao.txt >> .body

  TAO_COUNTER=$(($TAO_COUNTER +1));
  echo $TAO_COUNTER > .taocounter

  echo '-' >> .body
  echo ' ' >> .body
  echo 'The first version is from the Fortune files. The second version is the Beatrice Tao.' >> .body

  # jlj -bf .body -ec daily_tao -es `cat .taocounter` -ne -s
  cat .body | ./clive -u username -w password -j daily_tao -s `cat .taocounter` --charset utf8

I saw this, and it didn't make a ton of sense. Thankfully it's a very short script so I was able to piece it together quickly. Today I probably wouldn't write it this way, but… It works, so why would I touch it?

The new script is in Ruby. I say new because I learned Ruby after Bash, and once I knew Ruby, it was my preferred language. I'm guessing I wanted to rewrite the script so it would be more legible, and then… For some reason, I never replaced it.

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
  #!/usr/bin/ruby

  Dir.chdir '/home/user/www/tao'

  # Arrays go from 0... infinity
  # Tao goes from 1 to 81
  # Counter therefore goes from 0 to 80 and loops.

  counter = File.open('.taocounter', 'r') { |f| f.readline.strip.to_i }

  tao = File.open('tao.txt', 'r') { |f| f.readlines('%') }
  beatrice = File.open('beatricetao.txt', 'r') { |f| f.readlines('%') }
  disclaimer="The first version is from the Fortune files - a program for random quotes on GNU/Linux.\n The second version is the Beatrice Tao, available online, reproduced here with permission.."

  File.open('.body', 'w') { |f| f << "#{tao[counter]}\n\n________________\n\n#{beatrice[counter]}\n\n#{disclaimer}" }

  `cat .body | ./clive -u username -w password -j daily_tao -s "Tao Te Ching: Chapter #{counter+1}" --charset utf8`
  counter = 0 if (counter +=1 ) > 80
  File.open('.taocounter', 'w') { |f| f << counter }

My favorite thing here is probably the second-to-last line of the Ruby script.

There are no tests for any of this, because I definitely wrote this before I knew you could test your code.

And the question remains: should I write tests? Should I refactor? Should I fix it, rewrite it, etc?

And the answer remains: No. It works. There are no new features to add. Just leave it be.