When you receive Japanese text in an unfamiliar encoding, do not choose UTF-8 or Shift_JIS by intuition and overwrite the source. Inspect separate copies, record the detected encoding, compare a few diagnostic characters, and save any conversion under a new name. A Japanese text editor is useful here only if it exposes those decisions clearly.
Choose capabilities, not country of origin
A Japanese-made editor may offer a familiar interface or local support, but origin does not prove encoding coverage. Likewise, software made elsewhere may handle Shift_JIS well. Ask concrete questions instead: Does it display the detected encoding? Can it distinguish UTF-8, Shift_JIS, and EUC-JP? Does it report characters that cannot be represented in the destination encoding? Can you preserve the source while testing a conversion?
Those answers matter more than a “Japanese-friendly” label.
Compare the same text in three source files
The sample plan used the same 80 lines in three files and included two circled numbers and two wave dashes. The recorded inspections returned:
| Source | Detected value | Characters | Lines | Unconvertible candidates | Visual symbol comparison |
|---|---|---|---|---|---|
| UTF-8 | UTF-8 | 1,591 | 81 | 0 | not recorded |
| Shift_JIS | Shift_JIS | 1,591 | 81 | 0 | not recorded |
| EUC-JP | EUC-JP | 1,591 | 81 | 0 | not recorded |
The reader reported 81 lines because the 80 content lines ended with a final line break. Each file returned the intended encoding name and an empty unconvertible list. That is useful evidence about detection, but it is not a visual result for the circled numbers or wave dashes. “Zero candidates” does not prove that every glyph looked identical.
Pair a numeric check with a human check
First compare encoding name, line count, character count, and candidate count. Then place the four diagnostic symbols side by side and look for substitutions, a changed wave-dash shape, or missing characters. You need both checks: correct detection can coexist with a display difference, and a plausible display can still contain a character that fails in the destination encoding.
If you convert a file, write converted-utf8.txt instead of replacing the source. Reopen it, compare the four symbols, and confirm the count. Never preserve a deliberately mis-decoded view by saving over the original.
The bounded Rune Studio result
Rune Studio’s current Mac documentation describes detection of major Japanese encodings, marking of unconvertible characters, and pausing autosave while a conversion problem remains. In the recorded three-file run, UTF-8, Shift_JIS, and EUC-JP were returned separately; each source had 1,591 characters, 81 lines, and no reported candidates.
The run did not overwrite a mis-decoded source, visually compare the two circled numbers and two wave dashes, repair damaged data, perform OCR, or fix a binary file. Test the actual characters from your archive before migrating a large collection.
Stop guessing before you save
The safe decision is source-specific: identify the encoding, inspect diagnostic characters, convert only a copy, and reopen the result. Product origin can help with interface preference, but verified encoding behavior is the criterion that protects Japanese text.
Current details about the Mac app’s Japanese encoding support are available on the Rune Studio website.