I have recently read this article “Don’t Make Me Remember Things” where the author complains about string’s indexOf method having less-than-ideal API. Specifically, about the case when the sought substring is not found: indexOf usually returns a nonsense value like null or -1, and the developer has to remember that, which leads to bugs.

On second thought though, it seems like there are two problems mixed in this story:

  1. whether the substring is found at all,
  2. what is the exact position.

And if I look at indexOf this way, I would probably have 2 variables, explicitly covering for both of those purposes:

var index = string.indexOf(substring);
var isFound = index !== -1;

if (isFound) doSomethingAbout(index);
else doSomethingElse();

and in this case, there is no problem with forgetting.

To that this idea a littl further, I would probably have a different function to check the inclusion:

function containsSubstring(string, substring) {
  return string.indexOf(substring) !== substring;
}

if (containsSubstring(string, substring)) {
  var index = string.indexOf(substring);

  doSomethingAbout(index);
} else {
  doSomethingElse();
}

This version is calling indexOf twice, and there can be cases where this is a performance issue, and then I’d stick with the first approach.

* * *

It’s another case where an issue is unconfortable to deal with because it’s actually more issues mixed together, and, which are simpler when taken apart.