XML ↔ JSON Converter

A bidirectional converter that keeps what most XML-to-JSON tools throw away: attributes, mixed content and repeated elements. Attributes become keys under a configurable prefix, so <book id="1"> yields {"@id": "1"}; text sitting alongside attributes goes into a #text key; repeated child elements collapse into an array. The XML side is parsed by the browser's own DOMParser, the same code path used for an application/xml response, so namespaces, entities and CDATA behave exactly as they would in your page.

The conversion is inherently lossy in one direction because the two formats disagree about what a document is. JSON has arrays; XML does not, and infers them from repetition — which means the shape of the JSON depends on the data rather than the schema. One <item> produces an object, two produce an array of objects, and code written against a sample with two items breaks on the day a customer sends one. If a consumer needs a stable shape, normalise after converting rather than trusting the input to stay plural.

Ordering is the other casualty. XML preserves the sequence of everything, including the interleaving of text and elements in mixed content like <p>Hello <b>world</b> again</p>. JSON objects have no meaningful order, so that sentence cannot be reconstructed from keys alone. For document-shaped XML — anything where the elements are inside prose rather than around data — a JSON representation is a lossy summary rather than a translation.

Namespaces carry a subtler trap. The prefix in x:title is arbitrary and local to the document; the URI it is bound to is the part that identifies the vocabulary. Flattening to JSON keeps the prefix as text in the key and drops the binding, so two files describing the same thing with different prefixes produce different JSON, and two files using the same prefix for different vocabularies produce colliding JSON. If namespaces matter to what you are building, resolve them before the conversion, not after.