9
\$\begingroup\$

The task is to write a JavaScript-function which translates a given blank-separated sentence to title-case.

Means that all words shall start with a capital and then the rest of the word in lower-case. But: A certain, specified set of conjunctions, preposition as well as article shall be all lower-case.

Example: "The second of the four items." becomes "The Second of the Four Items.".

Here's my implementation of such a function:

function translateToTitleCase(str) {
  const translateWord = (sWord) => {
    return sWord.slice(0, 1).toUpperCase() + sWord.slice(1).toLowerCase();
  }

  const words = str.split(" ");

  words[0] = translateWord(words[0]);

  for (let i = 1; i < words.length; i++) {
    if (!["of", "and", "the", "to"].includes(words[i].toLowerCase())) {
      words[i] = translateWord(words[i]);
    } else {
      words[i] = words[i].toLowerCase(); // Make sure is's the correct case, when the sentence (or parts of it) is given in uppercase.
    }
  }

  return words.join(" ");
}

// -- Examples -------------------------------------------------------------------------------
console.log(translateToTitleCase("Into unmerciful the entreating stronger to of word guessing."));
console.log(translateToTitleCase("the OLD MAN aND THE sEa"));

I think my coding is still a bit "noisy" with the usage of all those brackets, chained methods and concatenation.

Any ideas about how to improve my implementation?

Perhaps some cool new ES6-feature I wasn't aware of.

What would you have done differently and why?

\$\endgroup\$
  • \$\begingroup\$ I'd argue that the functionality should be changed slightly – this'll trip over words like xkcd and eBay – but I suppose that's not really in-scope for Code Review. \$\endgroup\$ – wizzwizz4 Apr 7 at 16:53
11
\$\begingroup\$

I find this code quite reasonable on the whole, but you might consider the following suggestions relating to consistency, succinctness and semantics.


Move and rename the inner helper function

translateWord is reusable as a general utility function and seems better-placed in the global scope and renamed to titleCase. This meshes with similarly-named Ruby, Python and PHP builtins (titlecase, title, ucfirst, respectively). When I see "translate", I think of linguistics or mathematics before I think of strings or casing.

Avoid excessive calls to toLowerCase

.toLowerCase() is more efficient called once on the entire sentence before splitting rather than incurring the overhead of calling it multiple times per word. With this in mind, you can skip calling the titleCase function described above if you wish.

Improve "ignore" list

if (!["of", "and", "the", "to"].includes(words[i].toLowerCase())) {

is problematic for a few reasons:

  • It creates a new array object for every word. Move initialization to the top of the function and create it once.
  • Hardcoding restricts your function's reusability. Making this ignore list a default parameter allows the client to adjust the list as needed.
  • Giving this array a variable name makes its purpose more obvious.
  • Although the array is small, it needs to be traversed linearly to perform a lookup; using a set improves semantics, readability and time complexity all at once and is the ideal structure for testing membership.

Avoid the loop

This task is a map operation: each word has a function applied to it. You can roll split, map (your for loop) and join into one call to replace, which takes a regular expression that splits on non-word characters and applies the titleCase function to each one that passes the ignore test.

Minor points

  • sWord.slice(0, 1) can be sWord[0].
  • sWord is an okay variable name, but str (matching your outer function) or word seems more consistent.
  • Unless there is a good hoisting or context reason, I'd make the outer function also use an arrow function for consistency with your inner function.

A rewrite

const titleCaseWords = (words, ignore=["of", "and", "the", "to"]) => {
  ignore = new Set(ignore);
  
  return words.replace(/\w+/g, (word, i) => {
    word = word.toLowerCase();
    
    if (i && ignore.has(word)) {
      return word;
    }
    
    return word[0].toUpperCase() + word.slice(1);
  });
};

[
  "Into unmerciful the entreating stronger to of word guessing.",
  "the OLD MAN aND THE sEa"
].forEach(test => console.log(titleCaseWords(test)));

\$\endgroup\$
2
\$\begingroup\$
    if (!["of", "and", "the", "to"].includes(words[i].toLowerCase())) {

Not a coding issue exactly, but this is missing a lot of words that should not be capitalized. In particular, you only have one of the seven coordinating conjunctions, one of three articles, and two prepositions. Grammar Girl suggests

“a,” “an,” “and,” “at,” “but,” “by,” “for,” “in,” “nor,” “of,” “on,” “or,” "out," “so,” “the,” “to,” “up,” and “yet.”

She says this is from AP style (which the Associated Press apparently no longer uses).

Incidentally, you should also be capitalizing the last word of the title (as well as the first). Your two examples have nouns as the last word, so this gets capitalized correctly. You may want to add an example where the word is in the lower case list. E.g. "If it's the last word of the title, should I capitalize 'the'?"

If you were trying to implement the Chicago style, this would be much more complicated, as there are some words that can be prepositions, adjectives, or adverbs depending on usage. In Chicago style, you have to parse the sentence grammar to determine what gets capitalized. Also, Chicago style uses lower case for prepositions with four letters or more. So there are far more of them.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.