[qu4z-2]

I find

  map f xs = [f x | x <- xs]

the most readable, but given that list comprehension is basically a map and a filter joined together, that definition is kind of cheating.

I find the python version hardest to read (even though it's also "cheating"), which is largely because both the identifiers and the control constructs are alphabetic.

  [func(x) for x in iterable]

I have to read the words to figure out that for/in are the keywords and x/iterable are identifiers. func(x) is at least pretty obviously a function application. I'm glad it's not

  [call func x for x in iterable]

If you compare this to

  [f x | x <- iterable]

The parentheses-less function application might take some getting used to, but then it's pretty easy in my opinion. The | nicely divides it into two parts. A function application on the left, and a "take each element x out of iterable" on the right.

The only other thing is that because "iterable" is such a long word I expect it to be an identifier imported from a library or somewhere else in the program, and certainly not earlier in the line.

In summary, I think a lot of what we find "readable" depends on what we are used to.

[dchichkov]

>> In summary, I think a lot of what we find "readable" depends on what we are used to.

I think a general rule is that source code readability is inverse to the amount of context information that one needs to remember in order to interpret and understand the code.

If you can look at the code at any place (and at any scale, ranging from a single line to the whole call graph), and without any context understand what is happening at this particular place - code is readable. And on the opposite, if you need to know a lot about the context of each particular line of code - code is unreadable.