why not a,b or x,y?

It depends. x and y are either elements or coordinates, a and b usually elements though in e.g. Haskell reserved (by convention) for type variables.

The i j k l series is reserved for indices. n m etc. are the counts of something, as such you'll see i counting up to n. Both are due to mathematical sum notation and general mathematical convention. Random google result:

Let x₁, x₂, x₃, …x_n denote a set of n numbers. x₁ is the first number in the set. x_i represents the ith number in the set.

...if you're using a language in which you use i often chances are you should stop coding in C and get yourself a language with iterators. Manual loops are a bug magnet.