Character counts can answer different questions. A raw file count measures stored characters, while an editorial count may weight full-width characters as 1 and half-width characters as 0.5. In the controlled run, a specimen with a raw count of 24 displayed 13.5 characters in Rune Studio’s status bar. The difference is expected, not an arithmetic failure.
This article begins with the small sample, connects the visible UI value to its saved screenshot, and then assigns the raw CLI result to a separate use case.
Calculate the mixed-width specimen before opening the editor
The input had three lines: 全角ＡＢＣ, ASCII 123, and かなABC12. It combines Japanese and full-width Latin characters with half-width Latin letters, digits, a space, and line boundaries. The file also has a final newline.
Before counting a production manuscript, identify the recipient’s unit. “Characters” may mean a full-width/half-width editorial conversion, Unicode code points, grapheme clusters, bytes, or a platform-specific limit. A short known specimen makes the selected interpretation observable.
The acceptance value for this test is the editor’s actual weighted display. Line breaks and selection boundaries can make a hand calculation ambiguous, so the displayed 13.5 is recorded together with the exact input rather than generalized into an undocumented rule for every symbol.
In a hand worksheet, place full-width characters, half-width characters, spaces, and boundaries in separate rows. A total without categories cannot reveal whether a half-width space was included or a newline was excluded. When the worksheet and screen differ, review those categories and the selection before changing a coefficient.
Measure the selected range in Rune Studio
The run created an isolated workspace and opened mixed-width.txt with rune tab open. It selected the complete sample using rune tab set-cursor --offset 0 --length 24, then captured the Japanese main.single view. The persistent output is group1a_cycle5/STUDIO-193/main-single-selected-ja.png.
Visual inspection at original resolution shows the status bar reading 全体：13.5文字（選択範囲：13.5文字）, meaning total 13.5 characters and selected range 13.5 characters. Matching values establish that the complete measured sample was selected. This number was observed in the local product screen; it was not inferred solely from documentation.
If the selection value differs from the total when a whole-file measurement is intended, check the active tab, offset, length, and highlighted range. A cursor without a selection or a missing final character can silently change the result. Capture the screen only after the file identity and both status values agree.
For a partial production selection, record short quotations at both boundaries as well as the offset. Earlier edits can move an offset even when the intended passage is unchanged. A useful screenshot includes the filename, visible selection, and status bar in one frame.
Keep the raw 24 as a separately named measurement
An independent rune inspect text --path <mixed-width.txt> call reported characterCount=24, lineCount=4, UTF-8, and LF. This raw 24 describes the saved text inspection. The weighted 13.5 describes the editor’s full-width/half-width display. Neither should overwrite the other.
Raw values are useful for file identity, slices, and technical limits that define their unit that way. A weighted character count is useful when an editorial or commercial rule explicitly assigns half weight to half-width text. Put them in separate fields such as weighted=13.5 and raw=24 so a later reader cannot compare unlabelled numbers.
The CLI result also supplies a readback method. If the screenshot was taken from an unsaved buffer or the wrong file, its raw identity could differ from the named destination. Connecting the screenshot, path, encoding, and raw count keeps the measurement reproducible.
The reported four lines are supporting context. The sample contains three visible lines and a terminal newline, so another editor may present a visible total of three. Recording final-newline state prevents that display difference from being misdiagnosed as missing text.
Declare which number the recipient accepts
Use the weighted display for a submission rule, advertisement slot, or fee schedule that explicitly defines full-width as 1 and half-width as 0.5. Use a raw or other technical count when the API, conversion tool, or validation specification names that measure. If a rule does not state how spaces, newlines, emoji, or combining marks are treated, ask the rule owner instead of inventing weights.
For this specimen, acceptance requires the identified path and range, a visible total of 13.5, a visible selection value of 13.5, and a separate saved-file result of raw 24, UTF-8, and LF. The run does not establish weights for every symbol or emoji. Add a focused sample when a production manuscript introduces an untested category.
The practical value of text editor character counts comes from naming their units. Editorial planning can use the measured weighted value while engineering can use the raw file result, with both teams referring to the same artifact and neither claiming that its count answers every question.
If a deadline manuscript exceeds a limit, confirm the rule’s unit before deleting prose. Do not reinterpret raw 24 as weighted 13.5, or the reverse, merely to fit the threshold. Editing the passage is a separate approved action; the counting step should continue to report the selected rule faithfully.
Store the specimen beside the measurement record so the rule can be checked after a software update. If the same input later produces a different weighted result, repeat the screenshot and document the product version before applying the new value to a live commission. Historical counts should remain attached to the version that produced them.