An English-interface handoff for a Japanese manuscript should begin with an evidence-backed vocabulary card, not a generic “language switched” checkbox. Each card pairs an observed English screen term with the Japanese object it refers to, names the evidence, and assigns the next owner. Unknown labels stay blank instead of being guessed.
The retained screen provides one reliable anchor. It visibly combines the English heading Japanese Selection Boundary Test, the label Selected range:, and the selected Japanese text 東京. The image SHA-256 is a672e213695b2c79f34d40e72b9ade893f55fa740dd2aed6782dc89fb133a350. This proves one English-screen state containing Japanese text. It does not prove the act of switching the preference, complete menu localization, Japanese IME composition, or a saved manuscript round trip.
Hand off working vocabulary rather than a bare setting
An overseas operator needs to know which English cue to inspect, which Japanese object is in scope, and where to send a mismatch. A stored value such as en cannot answer those questions by itself. The handoff therefore uses task IDs: E-01 for the observed English cue, J-01 for manuscript integrity, and R-01 for environment restoration.
This article is not another preference round-trip test. Its judgment unit is a responsibility card anchored to visible evidence. The card can close one observed fact while leaving save behavior and the actual interface transition open for named operators.
Manuscript wording, encoding, IME behavior, and publication settings remain outside E-01. Stating those exclusions before the session prevents a localization check from quietly becoming an editorial change.
Anchor the first card to one visible English screen
At full size, the source image shows Japanese Selection Boundary Test as the heading and Selected range: 東京 in the work area. 東京 is visibly highlighted within mixed Japanese and Latin text. The tab name begins in English, while UTF-8 and LF appear in the lower status area.
Card E-01 records the observed term Selected range, the Japanese target 東京, the image path and digest, and the UI observer responsible for the result. An English-speaking operator can identify the cue, while a Japanese editor can identify the content under review.
The visible UTF-8 label is not evidence of a completed save-and-reopen encoding check. The visible Japanese characters are not evidence of IME composition. A screenshot records one state; it does not provide the preceding commands or subsequent saved bytes.
Split the handoff among three owners
E-01 belongs to the interface observer. Its input is the fixed image, its operation is full-size inspection, its output is the observed English term and Japanese target, and its acceptance condition is that both are legible in the same evidence frame.
J-01 belongs to the manuscript editor. Its future input is a Japanese source file with a recorded digest. The operator will use a disposable copy under the intended interface, save under a new name, reopen it, and compare content and digest expectations. That exercise is not present in the retained screenshot, so J-01 remains open.
R-01 belongs to the environment administrator. Its future input is the initial interface language and a recovery backup. The operator will switch the target profile, fully restart the application, record the resulting screen, and restore the original language afterward. The current image does not establish that transition, so R-01 also remains open.
Three cards make an incomplete test honest. A single green “English UI” status would hide which owner never performed a required step.
Grow the glossary only from observed terms
The initial confirmed glossary contains Selected range. Familiar words such as Search, Save, and Reopen may sound obvious, but they should not enter the confirmed column until the target version displays them. Put predicted terms in a pending column with the screen name, application version, and evidence still required.
Apply the same discipline to Japanese content. This screen establishes that 東京 is visible and selected. It does not establish treatment of every kanji, kana sequence, punctuation mark, or composition state. One exact anchor is more useful than a broad compatibility claim that nobody can reproduce.
In the final operator guide, describe the purpose beside each observed label rather than relying only on pixel location. Layout may move between versions; the editorial responsibility should remain understandable.
Route a mismatch without editing the manuscript
If Selected range is absent on the target screen, do not choose the nearest English phrase and continue. Record the application version, language state, and restart condition, then reopen E-01. The recovery path is to correct the observation conditions, not alter Japanese prose.
If the Japanese target differs from 東京, verify the source image digest, tab, and heading before diagnosing language support. The wrong evidence file can look like a localization failure. If a saved-copy digest changes unexpectedly, route the problem to J-01 rather than having the UI observer repair content.
If the original interface cannot be restored, keep R-01 open and recover from its recorded baseline and backup. The three-owner structure prevents an English-label problem, a manuscript problem, and an environment problem from collapsing into one vague failure.
Complete the handoff with two deliberate open cards
Only E-01 can close from the current evidence: the English heading and Selected range: 東京 are visible, and the source digest matches. J-01 lacks a measured saved-file round trip. R-01 lacks a measured preference switch, full restart, menu survey, and restoration. Japanese IME and notifications are also unmeasured.
Completion here means closing the fact that is actually observed and handing the two open cards to their named owners with explicit inputs and acceptance conditions. The next operator begins with J-01 or R-01 instead of reconstructing what a screenshot was assumed to prove.
For an English text editor UI Japanese text workflow, the practical result is a bilingual evidence map with bounded claims. It lets an English-speaking operator navigate an observed cue while the Japanese editor retains authority over content and the administrator retains authority over the interface environment. That is a safer handoff than treating one setting readback as proof of the entire editing session.