{
  "schemaVersion": "1.0.0",
  "appId": "count-it",
  "title": "Count It",
  "version": "2026-08-29.1",
  "launchUrl": "https://count-it.backwerdrhythmshop.com/",
  "launchUrlStatus": "Live. Cloudflare, custom domain.",
  "integrationLevel": 2,
  "integrationLevelRationale": "Accepts a configured assignment link, scores objectively, and emits the universal result envelope (praxis.result.v0_1) — the condition this manifest previously named for Level 2, now met. Transmission is NOT part of the claim and never has been: the envelope is a format, the card renders from it, and nothing leaves the device. Scale Trail claims Level 2 on the same terms.",
  "pathway": "rhythm-reading-and-counting",
  "supportedActivityTypes": [
    "choose-the-count",
    "practice-reading"
  ],
  "configurableSettings": [
    "a",
    "level",
    "scope",
    "meter",
    "cells",
    "guide",
    "fb",
    "retry",
    "n",
    "pass",
    "seed",
    "sys"
  ],
  "lockableSettings": [
    "level",
    "scope",
    "meter",
    "cells",
    "guide",
    "fb",
    "retry",
    "n",
    "pass",
    "seed"
  ],
  "assignmentLink": "A teacher pins a round in the URL and posts it: `cells` names an explicit rhythm vocabulary by catalog id, `scope` chooses one beat or one measure, `meter` chooses 4/4 or 3/4, `guide` fixes the subdivision-guide policy, `fb` chooses whether the correct answer appears after each question or only at the end, `retry` chooses what trying again means, `n` and `pass` set the length and the goal, and `seed` makes every student's questions identical. `level` is a shorthand for a cumulative vocabulary and is superseded when `cells` is present. `a` is a display name and `sys` names the counting system.",
  "assignmentValidation": "Rejects loudly and never repairs: an unknown rhythm id, fewer than 2 rhythms after duplicates collapse, a question count outside 1–20, a pass mark the round cannot reach (measured against the round's real length, 5 when the link sets none), a full-measure round longer than the pool can fill without repeating, a meter this app does not read, an unrecognized feedback or retry setting, or a counting system this app does not teach each invalidate the whole link with a plain-language message. A rhythm pool with a rhythm missing teaches a different step, so dropping one silently would produce evidence for an assignment nobody set.",
  "roundLengthRule": "A full-measure round never repeats a measure, so a pool of k ONE-BEAT rhythms can fill at most k^(beats per bar) questions: two rhythms make 16 in 4/4 and 8 in 3/4. A pool holding a whole or half note makes FEWER, because one cell fills several beats, and the ceiling is computed from the pool's real spans rather than from its size. Asking for more is refused at the link rather than while the round is being built — the shape of failure that let a banner state one assignment while the student answered another.",
  "rhythmCellIds": [
    "quarter",
    "eighths",
    "eighth-rest",
    "rest-eighth",
    "three-rest-note",
    "rest-sixteenth-rest",
    "alternating-rests",
    "rest-two-rest",
    "dotted-eighth-sixteenth",
    "eighth-two",
    "two-eighth",
    "sixteenth-eighth-sixteenth",
    "sixteenths",
    "rest-three",
    "two-rest",
    "rest-two",
    "half",
    "half-rest",
    "whole"
  ],
  "spanningCellIds": [
    "half",
    "half-rest",
    "whole"
  ],
  "spanningCellBeats": {
    "half": 2,
    "half-rest": 2,
    "whole": 4
  },
  "spanningCellStatus": "Whole and half notes, and the half rest — the values the Notes & Rests poster teaches that a one-beat catalog could not express. They sound once, at the top of the span, and the beats underneath are silent because the note is held: this app counts the notes that sound, so a half note on beat one of 4/4 answers \"1\". They are measure-scope only, and a beat-scope link naming one is refused. There is no whole rest: it fills the bar, so a measure containing one contains nothing else and has no count to ask for.",
  "meters": [
    "4-4",
    "3-4"
  ],
  "meterStatus": "4/4 and 3/4. A meter is chosen per link and defaults to 4/4, so every assignment written before meters existed means what it meant and generates the identical round. 3/4 reuses the whole vocabulary unchanged — same quarter-note beat, one fewer of them per bar. 3/8 was supported from 2026-08-24.1 to 2026-08-29.1 and then removed toward an app of its own: a link naming meter=3-8, or any of the retired eighth-beat cell ids (eighth-beat, two-sixteenths, sixteenth-rest, rest-sixteenth), is refused with a plain-language message rather than repaired.",
  "levels": [
    "level-1",
    "level-2",
    "level-3"
  ],
  "countingSystems": [
    "standard"
  ],
  "countingSystemStatus": "Standard American counting (1 e & a) is the only system this release teaches. The data model carries Eastman and Takadimi, but a link asking for either is refused rather than graded against a system the app does not present.",
  "acceptedInputSources": [
    "touch",
    "mouse",
    "computer-keyboard"
  ],
  "evidenceTypes": [
    "A1_ANSWER_CORRECTNESS"
  ],
  "compatibleSkillDomains": [],
  "skillVocabularyStatus": "Not yet reconciled. Candidate ids exist in the Sequence 2 draft but the vocabulary is an owner decision, and an invented id would be a contract no one agreed to. Until it is settled this app emits no skill ids.",
  "resultSchemaVersion": "praxis.result.v0_1",
  "resultSchemaStatus": "Adopted 2026-08-08, last of the three apps, after Scale Trail and Mallet Map. This app had no result object at all — the card was assembled inline — so the envelope was built rather than renamed. `skillReferences` is null and says so: the field is nullable across the family precisely so an app with no reconciled vocabulary can decline rather than invent ids. `activity.contentVersion` is null because questions are generated from the conditions and a seed, which already reproduce the round. In measure scope a miss is attributed to every cell in the measure, because the question is answered as a whole — the finest attribution the format allows, not a claim that all four were misread. Emitting the shape is not transmitting it: nothing leaves the device.",
  "supportDimensions": "The subdivision guide is a support, not a preference: an assignment pins it, and a gate counts the assignment's policy rather than the learner's own toggle. Question size (one beat or one measure) and rhythm vocabulary are the other two difficulty dimensions. There is no timer anywhere in this app, by design.",
  "integrityMeasures": "An assigned round varies the ORDER of the answer choices per student, from the seed plus an identifier, so the questions stay identical and comparable while a posted answer key does not transfer. Attempts are counted and reported on the card, in the copied summary and in the verification code, because a replay of an already-answered round is a different fact than a first attempt. Attempts are counted per assignment in browser storage, so the count survives a reload rather than resetting with the page. `fb=end` withholds the correct answer, the running score and the guide's highlighting until the round is over. `retry=reseed` gives a retake the same conditions on new questions, seeded from the link's own seed plus the attempt number so a teacher can regenerate any attempt; `retry=off` withdraws the retry control. Practice mode is closed while an assigned round is unfinished, since it reveals counts for the same vocabulary. None of this makes a score proof: the app has no accounts, `retry=off` cannot stop a page reload — it reports the further attempt rather than preventing it, which is the same posture this app takes toward a pass mark — and the verification code remains a deterrent rather than a signature.",
  "accessibility": [
    "keyboard-operation",
    "screen-reader-labels",
    "visible-focus",
    "reduced-motion",
    "large-touch-targets",
    "non-color-status-cues"
  ],
  "offlineBehavior": "No account, backend, or network dependency. Preferences, personal bests, a per-assignment attempt tally and an opaque per-browser string used only to vary answer order stay in browser storage; the optional assignment identifier is session-only and never persisted. None of the stored values name a person: the attempt tally is keyed by the assignment's own canonical link, and the ordering string is random and meaningless outside this browser. The visit counter is progressive enhancement and its absence changes nothing.",
  "siblingApps": {
    "mallet-map": "https://mallet-map.backwerdrhythmshop.com/",
    "scale-trail": "https://scale-trail.backwerdrhythmshop.com/"
  },
  "teachingSequence": "https://apps.backwerdrhythmshop.com/sequences/counting-rhythms/",
  "limitations": [
    "Whole, half, quarter, eighth and sixteenth values in 4/4 and 3/4. A whole note needs four beats and so appears in 4/4 only; a half note needs two and appears in 4/4 and 3/4.",
    "No triplets, compound meter, or ties across beats.",
    "No eighth-note-beat meters: 3/8 was removed in 2026-08-29.1 and is planned as its own app.",
    "Does not play, listen to, or time anything: there is no audio, no microphone, and no tempo engine.",
    "Does not measure live performance, tone, sticking, or physical technique.",
    "Standard American counting only in this release.",
    "Progress is device-local; no account, roster, or cross-device sync.",
    "The verification code is a deterrent, not proof — signed verification is future work."
  ]
}