Readable indentation should not quietly rename namespaces, reorder attributes, convert CDATA, remove comments, or rewrite an entity spelling. XMLClear keeps those constructs as source tokens after syntax succeeds.
Prefixes are spelling, namespace URIs are identity
Two documents may use different prefixes for the same namespace URI, and a default namespace changes the namespace of unprefixed element names within its scope. A formatter does not need to invent or normalize prefixes. XMLClear preserves every qualified name and xmlns attribute exactly in the source order and reports only how many namespace declarations the parsed tree contains. That count does not prove that the vocabulary is expected or that an application will bind it to the intended code.
Keep non-element nodes visible
Comments can carry operational notes, processing instructions can tell a consumer about an associated stylesheet or tool, and CDATA can make literal markup characters easier to author. XMLClear preserves their complete source spellings and shows separate counts. It never executes an instruction, renders SVG, applies XSLT, opens a referenced stylesheet, or inserts parsed nodes into the active page. Preservation means only that the text remains in the derived document, not that its effect is safe or understood.
A namespace name is not a download instruction
In <x:item xmlns:x="urn:example:catalog"/>, the prefix x identifies the namespace declaration used by item. The URI names that namespace; it need not point to a downloadable schema. Preserving this declaration does not validate the element against an application vocabulary.