I build an EU e-invoicing API, and the uncomfortable part of that is that the code generating a document and the code validating it are the same codebase, written by the same person. If I get an assumption wrong, both sides are wrong the same way and every document passes. So I wired up the official CEN conformance corpus and then ran my output through KoSIT's validator — the actual jar German public-sector receivers run. It rejected my CII output on an XML Schema violation: I was emitting address fields out of the sequence order the UN/CEFACT schema requires. The bit I found genuinely interesting: Schematron runs on top of a document that's already assumed schema-valid, so my validator was faithfully checking business rules against XML that wasn't well-formed per the spec, and reporting "valid". No amount of running my own validator would ever have caught it. Same serializer feeds my Factur-X PDFs, so that shipped the same defect silently for months. Happy to answer anything about EN 16931, Schematron, or PDF/A-3 - it's a weirdly deep rabbit hole.

InvoiceHub

EN 16931, BL 2.1, UN/CEFACT CII, XRechnung 3.0, Factur-X

EN 16931, BL 2.1, UN/CEFACT CII, XRechnung 3.0, Factur-X