Rude Word Hunt

This post is inspired by a post a friend shared with me on Twitter. It shows a set of famous chocolate bar wrappers that have been cut and reglued together.



That immediately got me thinking if I could get a computer to do this for me. There are so many other situations where this could come in handy. For example, have you ever found yourself alone with a letterboard and wondered what fun you could make by rearranging the letters? Wonder no more.

There will be some rude words in this article. If you’re not too worried about that, read on.

The source code for this article can be found on Github.

Some Ins and Outs

The program takes a list of everyday words and returns a list of rude words.

Each rude word returned will have a pipe character inserted at points where an input word would have to be cut to form the rude word. For example, let’s say the program was given “twix” and “kit kat”. It would return “tw|at”. It would be up to the user to figure out from which input words the “tw” and the “at” came from.

Getting the Thrust of it

How does the matching work? We use a shrinking sliding window. In this section we’ll go through the concept, with examples.

The program will try to build each rude word one at a time from the list of given words. Let’s imagine that the program is trying to build the word “twat”.

The program starts by trying to find instances of the whole word “twat” in the given words. Unless the user had passed in one of the nine lucky English words that contain the letter sequence t-w-a-t within them, the program will keep looking.

The next step is key. The program will shrink the search word by one character and keep shrinking until it’s eventually just looking for the rude word’s individual letters. An example helps understand this.

• 4 letters: twat
• 3 letters: twa, wat
• 2 letters: tw, wa, at
• 1 letter: t, w, a, t

Let’s say we have our favourite input words “twix” and “kit kat”.

1. “twat” is not found in the input words, so try looking for something smaller. 
2. Neither “twa” nor “wat” are found, shrink the search words again. 
3. Aha! “tw” and “at” are found. 

This technique will give you longer fragments first, meaning there is less chopping to do if you are cutting the words out of a magazine or poster.

Massaging the Input Words

How do we stop parts of the input words being used multiple times for one rude word? We remove them from the list, that’s how.

Let’s say we have the input words “tap” and “water” and we need to remove “wat”.

First, we remove “water” from the list of input words, then add in any parts of that word which aren’t matched by the rude word fragment. In our example, we remove “water” and then add “er”. Thus, the list of words available for further matches will be “tap” and “er”.

It’s just like cutting letters out of a magazine for a ransom note.

Getting Busy With Some On The Side

There is one more key piece of this puzzle we need to go through for it to work - what happens when a portion of a rude word is matched?

Say we want to try build the word “threesome” and the program has just matched up the letters “eeso”. We still need to match “thr” and “me” before we can say the word threesome can be built from the input words.

The good news is that we can do this with very little effort... because we have already done most of it. All we need to do is to try build the rude word “thr” with the updated list of input words. Same again for “me”. Once we have all the parts we just put them together with the pipe separator. Word found.

This technique is another example of recursion which is quite commonly used where a problem can be divided into smaller and smaller parts.

Finishing Off

Once we have a list of rude words, we can order them in several different ways.

Some ways to order them are:

• By max fragment length from the input words - useful when you want to impressive that special someone with the longest chunk from the input words.
• By the longest rude word that can be formed - also useful when trying to use as many letters as possible.

If You Feel Up For Round Two

One way to take this little program further is to find multiple rude words for the input words and try to optimise for the most number of rude words, or the most letters used. To do that will require a little extra grunt.