Case Converter

How it works

1. 01

   Type or paste an identifier or a phrase. Spaces, underscores, hyphens, dots and slashes are all read the same way — as separators — and runs of them collapse to one boundary.

2. 02

   The splitter breaks the input into tokens at every separator and at every case boundary. parseHTML5String becomes parse, HTML5, String. The tokens are listed under the field, so an output that looks wrong can be traced to the split that caused it.

3. 03

   All twelve conventions are built from that one token list. Only the joiner and the capitalisation differ, which is why they can never disagree with each other.

4. 04

   Turn off "Keep acronyms uppercase" to get XmlHttpRequest instead of XMLHttpRequest. The toggle only touches the rows that capitalise words — snake_case is lowercase by definition and SCREAMING_SNAKE_CASE is uppercase by definition.

5. 05

   Copy any row with the button beside it. Everything runs in the page: nothing is uploaded, and nothing is saved between visits.

How the splitter finds a word boundary

Two rules, applied over Unicode letters, marks and digits. First: a boundary goes wherever a character that is not an uppercase letter is followed by one — that catches fooBar and md5Hash. Second: an uppercase run ends one character early when its last capital starts a lowercase word, which is the acronym rule and is what turns XMLHttpRequest into XML, Http, Request instead of X, M, L, Http, Request. Digits attach to the token in front of them, so HTML5, utf8mb4 and v2 stay whole. Anything that is not a letter, mark or digit is a separator and is dropped.

parseHTML5String → parse | HTML5 | String → parse_html5_string

Use cases

Questions

How does it split XMLHttpRequest?

Into XML, Http and Request. An uppercase run ends one character early when its last capital begins a lowercase word. Without that rule the same input would come out as X, M, L, Http, Request.

What happens to digits?

They stay attached to the token in front of them, so HTML5, md5, utf8mb4 and v2 each stay whole. A digit followed by a capital is a boundary, which is why parseHTML5String splits into parse, HTML5, String rather than parse, HTML, 5, String.

Why is iOS split into i and OS?

Because a lowercase letter followed by a capital is always a boundary, and nothing in the characters says iOS is one word. IPv6Address splits as I, Pv6, Address for the same reason. Recognising either would need a dictionary of brand names, and a dictionary that is wrong about your identifier is worse than a rule you can predict.

Where did my punctuation go?

It was treated as a separator. Everything that is not a Unicode letter, mark or digit splits tokens and is then dropped, so pasted prose loses its commas and full stops. Apostrophes are the one exception: one sitting between two letters stays inside the token, so a word survives whole in the prose rows and is stripped in the identifier rows.

What does "Keep acronyms uppercase" do?

It leaves a token that is entirely uppercase — HTTP, XML, ID, HTML5 — spelled that way in the rows that capitalise words: PascalCase, Train-Case, Sentence case, Title Case, and camelCase after the first token. snake_case, kebab-case, dot.case and path/case are lowercase by definition and SCREAMING_SNAKE_CASE is uppercase by definition, so the toggle does not reach them.

Which words stay lowercase in Title Case?

Twenty-eight of them: the articles, the coordinating conjunctions, and prepositions of four letters or fewer — a, an, the, and, but, or, nor, for, of, on, in, to, with, from, into, over among them. All are capitalised anyway when they are the first or last token. That is the Chicago convention; AP capitalises anything four letters or longer, so the two disagree on with and from.

Is anything uploaded?

No. The splitter and all twelve conversions run in the page. Nothing is sent anywhere, nothing is logged, and nothing is stored between visits — reloading clears the field.