Behavior and limitations

Digits and word boundaries

Digits are not treated as word separators. In automatic detection, digits are also treated like lowercase characters by default when checking whether an input consists only of uppercase letters. This can make strings containing digits split differently than expected:

use Jawira\CaseConverter\Convert;

(new Convert('REEL2REAL'))->toDot(); // r.e.e.l2.r.e.a.l
(new Convert('reel2real'))->toDot(); // reel2real
(new Convert('Reel2Real'))->toDot(); // reel2.real

For example, hello8World converts to hello8-world in kebab case. The converter cannot infer that the 8 should have been a separate word. If a particular split is required, preprocess the input to add a separator where needed.

Acronyms and mixed capitalization

The converter splits before every uppercase letter in unseparated input. It does not infer acronym boundaries.

For example, XMLHttpRequest is split as X, M, L, Http, Request, and converts to x.m.l.http.request in dot notation. If the intended words are Xml, Http, and Request, write the input as XmlHttpRequest.

❌ Wrong	✅ Good
PHP_CodeSniffer	php_code_sniffer
PHP_CodeSniffer	PhpCodeSniffer
PHPUnit	PhpUnit
PHPUnit	php-unit

Use consistent separators and capitalization in the source. For example, PHP_CodeSniffer mixes an acronym with an underscore-separated name, while php_code_sniffer and PhpCodeSniffer have unambiguous boundaries.

Acronyms in PHP

This acronym problem has also affected PHP's own class naming conventions. A 2024 PHP RFC on acronym casing was accepted to treat acronyms and initialisms like regular words in Pascal case: for example, prefer HttpRequest over HTTPRequest. The RFC updates PHP's policy for internal classes; it does not rename existing classes. This mixed-case style also makes word boundaries easier for Case Converter to detect.