# Scope and limitations

These 36 constructed cases document the pinned conversion implementation, not the frequency of encoding errors or every browser. Byte preservation means exact byte equality after decode and re-encode. It does not mean semantic equivalence, identical rendering, or successful recovery of the original file encoding. Invalid UTF-8 triggers the site’s whole-input byte-to-Latin-1 mapping; re-encoding that text as UTF-8 changes non-ASCII bytes. Initial UTF-8 BOM removal can also change bytes for valid UTF-8. No normalization is performed. Automatic direction detection, 7-bit decoding heuristics, lone JavaScript surrogates and large-input limits are outside this casebook. No visitor data is used.