Identify the Likely Legacy Encoding Before Converting

Before selecting a conversion profile, find out which legacy font or keyboard layout created the source text. Ask the original author, inspect the document's font list, or look for documentation from the software that produced it. The same language may have several legacy encodings, and a mapping that works for one font can scramble another. A few visible words are not enough to identify the source reliably; include a sample with common letters, vowel signs, punctuation, numerals, and combined forms.

Choose a short sample that appears in a trusted original rendering and convert only that sample first. Compare every character, not just whether the output looks generally readable. Check names, dates, addresses, and technical terms separately because a small mapping error can change their meaning. If the converter offers several profiles, test each plausible one against the same sample and retain the profile that matches the source consistently. Do not select a profile simply because it produces the most familiar-looking line.

Once the sample passes, save a copy of the original and convert a small batch from the same source. Open the output in another Unicode-capable editor and test searching, copying, and pasting. If the text remains correct outside the original application, that is evidence that the conversion produced standard characters rather than relying on a special font. Keep notes about the source font, selected profile, and any manual corrections so another person can reproduce the process.

Mixed-source documents need extra care. A file may contain sections typed with different fonts or pasted from different systems. Convert those sections separately when necessary, and review tables, footnotes, and names after the main body. The converter handles character mapping; it does not automatically repair page layout, spelling, or the historical meaning of a damaged document. For legal, academic, or archival records, have a fluent reader verify the converted text before treating it as authoritative.

A useful verification sample should include text from the middle and end of the document as well as its opening lines. Some legacy mappings appear correct for common characters but fail on punctuation or less frequent combinations. If a converted page contains names in multiple scripts, ask a reader familiar with each script to review those passages. Record unresolved characters rather than silently replacing them with a guess, especially when the file is part of an archive or official record.