New Operation: Word Count

[neocotic]

Add a new operation that counts the words within its rendered contents.

[neocotic]

A few things to note on the new wordCount operation;

• It will simply count all character blocks that are separated by white space (i.e. spaces, tabs, new lines, or a mix of all three)
• It will not differentiate between what is a word or an HTML block etc. so it should be used on plain text for best results

Finally, I've been testing various methods for counting words and trying to find the best performing solution as this could become a popular tool (e.g. easily count words in the selected text).

I came up with 4 different approaches and tested them first to ensure they worked as expected: http://jsfiddle.net/alasdair/d3KGy/. Then I created a simple performance test to find the best solution and found some surprising results: http://jsperf.com/word-count-test.

Surprisingly, on Chrome at least, the approach with the most code performs best. This is best translation (I think) into CoffeeScript:

count   = 0
text    = text.trim()
matches = text.match /\s+/g

if text
  count++
  count+= matches.length if matches

count

[neocotic]

This has been implemented by PR neocotic/template-chrome#1. One main caveat that I'll possibly fix in the future: special characters are current considered as "words". For example; foo - bar would be counted as having 3 words, as the hyphen would be counted. This is due to the optimized and way in which we're counting words based on spaces between text and words themselves. Also, how am I to know what should be considered a word (e.g !@~# could be considered a - possibly censored - word by some).