Discovering Primitive Objects In JavaScript (Part 1) — Smashing Magazine

It seems natural to use strings to distinguish things. It’s very likely that in your codebase, there are objects with name, id, or label properties that are used to determine if an object is the one you’re looking for.

if (element.label === "title") { make_bold(element); }

At a certain point, your project grows (in size, importance, popularity, or all at once). It needs more strings as there are more things to distinguish from each other. The strings grow longer, as does the cost of typos or, say, your label naming convention changes. Now you have to find all the instances of those strings and replace them. Consequently, a commit for that change becomes much bigger than it should be. Which makes you look better in the eyes of the clueless. Simultaneously it makes your life miserable since it’s much harder now to find the cause of regression in your git history.

Strings are bad for identification. You have to consider uniqueness and typos; your editor or IDE won’t check if it’s the string you meant. It’s bad. I hear someone saying, “Just...