What is an ideal variable naming convention for loop variables? [closed]

Answer 1

@JustMike . . . A FEW C EXAMPLES: . . . to accompany the Java ones.

NON-NESTED loop: . . . limiting scope where possible

/*LOOP_DESCRIPTION*/ {

    int i;

    for (i = 0; i < LOOP_LENGTH; i++) {

        // loop body
    }  
}

NESTED loop: . . . ditto

/*LOOP_DESCRIPTION*/ {

    int row, column;

    for (row = 0; row < ROWS; row++) {

        for (column = 0; column < COLUMNS; column++) {

            // loop body
        }
    }  
}

One good thing about this layout is it reads badly without comments, thereby encouraging them.
It's verbose perhaps, but personally this is how I do loops in C.

Also: I did use "index" and "idx" when I started, but this usually got changed to "i" by my peers.

Answer 2

i also use the double-letter convention. ii, jj, kk. you can grep those and not come up with a bunch of unwanted matches.

i think using those letters, even though they're doubled, is the best way to go. it's a familiar convention, even with the doubling.

there's a lot to say for sticking with conventions. it makes things a lot more readable.

Answer 3

Whatever you choose, use the same index consistently in your code wherever it has the same meaning. For example, to walk through an array, you can use i, jj, kappa, whatever, but always do it the same way everywhere:

for (i = 0; i < count; i++) ...

The best practice is to make this part of the loop look the same throughout your code (including consistently using count as the limit), so that it becomes an idiom that you can skip over mentally in order to focus on the meat of the code, the body of the loop.

Similarly, if you're walking through an 2d array of pixels, for example, you might write

for (y = 0; y < height; y++)
  for (x = 0; x < width; x++)
    ...

Just do it the same way in every place that you write this type of loop.

You want your readers to be able to ignore the boring setup and see the brilliance of what you're doing in the actual loop.

Answer 4

Steve McConnell's Code Complete has, as usual, some excellent advice in this regard. The relevant pages (in the first edition anyway) are 340 and 341. Definitely advise anyone who's interested in improving their loop coding to give this a look. McConnell recommends meaningful loop counter names but people should read what he's got to say themselves rather than relying on my weak summary.

Answer 5

My favorite convention for looping over a matrix-like set is to use x+y as they are used in cartesian coordinates:

for x in width:
    for y in height:
        do_something_interesting(x,y)

Answer 6

If the counter is to be used as an index to a container, I use i, j, k.

If it is to be used to iterate over a range (or perform a set number of iterations), I often use n. Though, if nesting is required I'll usually revert to i, j, k.

In languages which provide a foreach-style construct, I usually write like this:

foreach widget in widgets do
  foo(widget)
end

I think some people will tell me off for naming widget so similarly to widgets, but I find it quite readable.