# "a" or "an" with big O notation?

A simple (though somewhat pedantic) question: when using big O notation in a sentence, should it be preceded with "a" or "an"? Example:

The extra loop results in a/an $\mathcal{O}(N)$ increase in cost.

The ubiquitous Cormen textbook notes on page 47:

For a given function $g(n)$, we denote by $\mathcal{O}(g(n))$ (pronounced “big-oh of g of n” or sometimes just “oh of g of n”)

which would seem to indicate "a" should precede something pronounced "big-oh". Later on the same page, however, the textbook uses the phrase "an $\mathcal{O}(n^2)$ upper bound." Which is used more in the literature? It seems like a triviality, but as someone working on the edge of theoretical computer science from a primary focus on another field, I'd like not to sound like an outsider.

• I think the common way to pronounce the offending sentence would have just “oh en increase” with no “big”, so an “an” is appropriate. Oct 8 '14 at 16:54
• Speaking of pedantry, the proper way to typeset the big-oh notation, as it has been used for over a century, is $O(\dots)$. (I observe that the Cormen textbook agrees, I don’t know why you changed it in the quote.) The caligraphic $\mathcal O$ just displays a lack of taste. Oct 8 '14 at 17:11
• (Tell me to shut up if I’m annoying.) While we are on the subject: on the other hand, a calligraphic $\mathcal P$ is an appropriate power-set notation. A depressingly large number of people abuse for that purpose the symbol for the Weierstraß elliptic function $\wp(z;\tau)$. Oct 8 '14 at 19:34
• I may be the only one, but when I am wondering about similar issues I mentally expand "$O(N)$ increase" to "order of $N$ increase". Oct 8 '14 at 20:33
• "...increases the running time by a factor of O(n)." (And pace Cormen, I've always pronounced it "order en".) Oct 12 '14 at 16:29