Relative URL reference

A Relative URL reference identifies a target using a base URL to supply missing context. It can contain a path, a query string or a fragment identifier, without independently specifying an entire network address. Resolution combines the reference with the base according to a defined algorithm.^[ref]

Trailing slashes provide an illuminating comparison. Resolving c against https://example.org/a/b gives https://example.org/a/c; resolving it against https://example.org/a/b/ gives https://example.org/a/b/c. In the first case the last base path segment is replaced, while in the second it names a directory-like prefix.

A parent-directory step is similarly concrete: ../image.png against https://example.org/a/b/page.html gives https://example.org/a/image.png. These outcomes follow from URL processing, not from examining an actual filesystem. No network request is needed to calculate them.^[ref]

The base is therefore part of the input. Copying the same relative spelling into a document with a different base can change its target even when every character of the reference stays unchanged.