C# algorithm for N-gram

I am intending to use the n-gram code from this article. The algorithm produces these tri-gram开发者_StackOverflow results:

t, th, the, he, e, q, qu, qui, uic, ick, ck, k, r, re, red, ed, d

for the text the quick red

However wikipedia, reckons it should be:

the  qui  k_r
he_  uic  _re
e_q  ick  red
_qu  ck_

(space indicated by ‘_’).

What is correct? Are there any other C# implementation out there?

The second example is correct.

ps. Why do you generate trigrams for the complete text and not only for words? What is your use case?

The first is correct. I uses character N-gram on my thesis. You must move forward and pass one character for each step. In this condition, similar words can be found.