XML Updating — Proof of Concept

Technical prototype portfolio — demonstration environment

Runs against the bundled sample XML document below, its own bundled real XSD schema, and the bundled update instructions below — some deliberately wrong — click Apply updates. No AI step — every update is targeted by a real XPath expression against the document's own real structure. A change that would break the document's own real schema is caught by a real validator (xmllint-wasm, a real libxml2 build) immediately after it's made and rolled back before it ever reaches the saved output — never applied and fixed up after the fact. An XPath matching zero or more than one real node is rejected rather than guessed at. Every applied update is then independently re-verified by re-parsing the real saved output from scratch, and every node no instruction targeted is confirmed to have genuinely kept its original value. This public demo only runs the bundled sample below — it does not accept an uploaded document or custom instructions.

Loading…

//caseRecord[@id='CASE-002']/status = Closed
//caseRecord[@id='CASE-003']/attorney = J. Rivera
//caseRecord[@id='CASE-001']/status = Dismissed
//caseRecord[@id='CASE-004']/docketNumber = ABC-123
//caseRecord[@id='CASE-999']/status = Closed
//caseRecord/status = Archived