A generic XML element: a name, an ordered attribute list, an ordered child list, and the source span it occupies.

Nothing here knows what any name means. The element is whatever the XML said - no validation, no namespace resolution - and giving it meaning is the lowering pass's job (docs/architecture.md, "Parser: DOM first, then lowering"). Nothing on the element itself is normalized either; the one normalization the parser does perform is XML 1.0 3.3.3 on attribute values, which an XML processor owes its application before a value exists at all (ADR-0043) - see Statifier.Parser.DOM.Attribute.

• name is the raw qualified name exactly as it appeared, prefix included.
• attributes are in document order, with duplicates preserved: a document that declares the same attribute twice is a thing the validator gets to report, so the parser may not silently drop one of them.
• children are in document order and interleave elements with Statifier.Parser.DOM.Text, so mixed markup is representable. Whitespace-only text is a child like any other; Statifier.Parser.DOM.elements/1 is the filter for readers that do not want it.
• location spans from the < of the start tag to the byte after the > of the end tag (for a self-closing element, the tag itself).