A repeated prompt is a review task, not an automatic mistake. Different classes, editions or contexts can give the same short question a different answer. CardQuire groups repeated question text and asks you to decide what the paper edition needs.
Work through the example
| Card | Source | Prompt / answer |
|---|---|---|
| C0004 | Printing basics · record 5 | Workshop start time? 09:00 — morning group. |
| C0008 | Workshop notes · record 5 | Workshop start time? 14:00 — afternoon group. |
For a morning-only handout, choose Keep only C0004. C0008 is excluded from print, but its answer and original row are retained. Open Sources to edit it and include it again.
For a shared handout, you could edit the first prompt to “Morning workshop start time?” and the second to “Afternoon workshop start time?”. They are now different prompts. Alternatively choose Keep all variants after deciding that identical wording is useful for your particular exercise.
Matching text is deliberately narrow
Question matching trims surrounding whitespace and uses Unicode NFC normalization. It is case-sensitive and does not rewrite internal spaces or punctuation. “What scale?” and “what scale?” are different; composed and decomposed spellings of the same accented character can group together.
The original strings stay unchanged. There is no semantic model, fuzzy answer judgment or factual correctness score. “Same question and answer” means a normalized text match, not that either answer is correct.
Decisions survive more chapters
Adding an unrelated chapter keeps your existing review decisions. Adding another card to a reviewed prompt creates a changed group, so it needs review again. Editing a member also reopens the decision. Reordering an otherwise unchanged group does not invalidate it.
Two uploaded files can have the same filename. They still receive separate source IDs and every card gets its own stable card ID. Selection is tied to the card ID, not a filename or the position of a button.
Edit where the card lives
Click a card on the paper. Change either side, then Save card. Move earlier or Move later saves the edits and changes its position in the retained edition sequence. Excluded cards still occupy that sequence; they do not take up print slots.
Open the original-record disclosure to compare the untouched cells, chapter label and source lines. Print slots may change after an edit or exclusion, but C0004 remains C0004. This makes it possible to discuss a correction without relying on a temporary page number.
Before printing
Check all included cards with empty fields. Resolve repeated prompts. Then read both faces, especially any question that lost its surrounding chapter context. A short flashcard may be accurate in the source chapter yet ambiguous when cut into a loose stack.
Continue with the front/back pairing and overflow checks. Save an edition JSON before making a different version for another class.
Read a card after removing the chapter around it.
Imagine a fictional print workshop with two groups: the morning group begins at 09:00 and has twelve places; the afternoon group begins at 14:00 and has eight. A chapter heading makes “When does it start?” understandable on a screen. After cutting the sheet into loose cards, that heading may be in a different pile.
The original ambiguous card set and revised card set use the same four invented facts. They are teaching fixtures, not a native Gemini Notebook export or a claim that the facts describe a real event.
- Before: “When does it start?” → “09:00.”
- After: “At what time does the fictional morning print workshop start?” → “09:00.” The question now names the group whose time is being recalled.
- Before: “How many places?” → “Eight.”
- After: “How many places are available in the fictional afternoon group?” → “Eight.” This preserves the count without letting the morning group’s twelve become an equally plausible answer.
- Before: “What is the later time?” → “14:00.”
- After: “At what time does the fictional afternoon print workshop start?” → “14:00.” “Later” depended on another card or paragraph; the revised question stands on its own.
- Before: “What is the other capacity?” → “Twelve.”
- After: “How many places are available in the fictional morning group?” → “Twelve.” Naming the group removes a hidden comparison without adding a new fact.
Edit the dependency, not just the wording.
Load the ambiguous set as one chapter. Map Question and Answer, and mark the first record as a header. Keep Context as source evidence rather than printing it automatically as the answer. Read each question with its chapter title covered. If more than one answer fits, decide which missing subject, date, version or condition is necessary.
Use the revised set as a comparison, or make the same edits in the edition. Do not include both versions merely to increase the number of cards. If you do import both, use the retained source records to explain which version belongs in the final paper set; exact repeated-prompt detection will not identify differently worded questions as the same concept.
Keep the answer small enough to inspect.
A longer question can carry essential context while leaving one checkable answer on the reverse. Conversely, splitting one answer into several cards can remove a condition that made it true. Compare the edited card with its source before excluding the original. The print layout can report overflow, but it cannot decide whether a rewrite changed meaning.
Finish by reading a shuffled paper preview without chapter order as a clue. This is an editorial review method, not a measured claim about learning performance. Google’s flashcard help describes native controls and CSV download; the wording and fictional workshop in this exercise are original.