for loops

Your code executes ls and records the results in a variable. The result is some text, a string of characters. (We call them “strings” and i is now a string variable.) Among the characters in a string variable might be spaces, tabs, or new line characters. I mention this because the special variable IFS is used by for loops, and it contains exactly one space, tab, and new line by default.

When you call for with a string as the input, it splits the string into units by splitting on each character in IFS. That is, it splits the big string into individual parts by splitting at each space, tab and new line. So this creates an array which is what is looped over. Each word in turn is assigned to your looping variable and then the code after the do is executed once per word.

(“word” has a sort of a special meaning here. When I say word, I mostly just mean a string that has no spaces in it. When you read text in English, there are words. They’re strings of characters separated by spaces. But words can also be separated by tabs, new lines, commas, semicolons, or whatever, but not by default when using for! You have to modify IFS to add those characters if you want them to be considered word separators.)

So, if any of the file system entries returned by ls have spaces in them, your loop is going to create more outputs than there are file system entries in the current directory.

For example:

file one.txt
file two.txt
my photo.jpg
notes (final).md
a b c d.txt

That would cause like 12 loops and 12 outputs in your code despite there only being five files.

If you instead overwrite IFS before running your loop, and only assign a single new line to the variable, then your loop will only be over the actual lines of the input text. Like this:

IFS=$'\n'

and then use your exact code above. Using my example of five files, this code will now only produce 5 outputs, not 12.

You can assign whatever characters you want to IFS.

(I have not tested any of this code, or examples.)

Variable names

The loop variable name i is just an identifier. Any valid variable name would work except you can’t use the reserved names like $1, $2 or any keywords as names. Also, there’s no way to escape an identifier. They are just literal names.

You also don’t want to use any built-in variable names or else you’ll overwrite their values for the duration of the current session. Bash will happily let you use them as your looping variable, but the rest of your code might have undesirable results. Variable names like IFS, for instance. :D