The inspiration for this extension is a neologism called Typoglycemia (http://en.wikipedia.org/wiki/Typoglycemia) which is the "purported recent discovery about the cognitive processes behind reading written text".

For example, most people can understand the following even though the text has been scrambled:

"I cdn'uolt blveiee taht I cluod aulaclty uesdnatnrd waht I was rdanieg"

The original text is:

"I couldn't believe that I could actually understand what I was reading"

Word Welter randomises the characters in all words in a webpage, except for the first and last characters.

The scrambling code has been developed so that it can cope with many different types of words e.g. words ending in punctuation, words with apostrophes. However, this is not a comprehensive text parser so please let me know if you find any issues. 

Unlike the example taken from the aforementioned Wikipedia page, for a word like "couldn't", I have chosen to fix the first character and also the last three (i.e. "n't") and then scramble all the rest; I found that the words get much harder to decipher if this approach is not taken!