Three consecutive guards and models were wrong in ways reading them would never
have shown, and each was caught by running it: the Room schema check that Room
itself defeated, the boundary check that examined nothing, and three modelling
faults in the engine. Recorded as a habit rather than three coincidences.
PersonalPredictionEngine keeps a discrete probability distribution over
candidate start dates rather than a date with a margin bolted on. Everything the
product needs falls out of that one structure: the most likely date is its mode,
the window is the narrowest span holding 80% of the mass, and a "Not yet" is the
distribution conditioned on what the user just said — which is what §13 asks for
and what a date-plus-margin design cannot express at all.
It is better, and that is a number rather than an opinion. EngineComparisonTest
scores both engines over the §51 fixtures on every build:
engine MAE mean window within +/-2 window covered
baseline 1.00 2.67 7/9 7/9
personal 0.67 4.56 9/9 9/9
COVERAGE IS THE MEASURE, NOT WIDTH
The first version of that test asserted the new windows must not be wider, and
it failed. Measuring showed why the assertion was wrong: the fixtures where the
personal engine is wider are the ones that are genuinely less certain — a
history with a suspected missing period, and one with a 45-day outlier — and the
baseline answers both with a two-day window and misses. What a window promises
is that the period starts inside it. An engine keeping that promise 7 times in 9
has a broken promise, not a tight forecast. The test now asserts coverage, with
a ceiling so "some time this month" still fails.
THREE MODELLING BUGS THE TESTS FOUND
Each was found by a test failing, not by reading the code:
- Median absolute deviation alone reads a user alternating 25 and 37 as
perfectly consistent, because half her deviations are zero. Twenty
disagreeing cycles came back High, breaking §15's rule that volume alone
must never buy High confidence. Spread is now the larger of MAD and mean
absolute deviation; robustness comes from IntervalAnalysis down-weighting
what is questionable, which is a better place for it.
- Recency weighting assumes the recent past predicts the near future. For a
variable user that is false — her latest cycle is a draw from a wide
distribution, not a signal — and weighting it equally cost three days on the
§51 variable fixture. Recency is now trusted in proportion to how much her
cycles actually agree.
- A fixed one-day floor on trend detection fired on a 42-day-cycle history
whose medians differed by a single day, turning an exact forecast into a
wrong one. One day is a real trend at 28 and rounding error at 42, so the
floor is relative to the user's own spread.
WIRED THROUGH, NOT JUST TESTED
PredictionInput carries recentAbsoluteErrors, and CycleRepository feeds the
scored errors back in. Without that the app stores every error it makes and
never reads one back — measuring accuracy rather than learning from it, with the
widening happening only in a unit test. A repository test asserts the errors
actually reach the engine.
BaselinePredictionEngine stays as the control, and both engines run the same
§51 acceptance suite, so the next engine's improvement is measurable too.
108 tests, all passing. ./gradlew check green. Verified on a device.
closes#10closes#11closes#12closes#14
IntervalAnalysis turns confirmed starts into weighted intervals and decides what
looks questionable — relative to this user's own history, never to a constant.
That distinction is the whole point. A global "over 40 days is suspicious" rule
gets exactly one group wrong, and it is the group whose cycles are already
unusual: the person this product exists for, and the one most tired of apps
assuming she is average. 45 days is unremarkable at a usual of 43 and worth
questioning at a usual of 29. Both are tests.
Two flags, because they earn different responses. A gap near a whole multiple of
the usual is a probable missed entry and produces §14's question. A gap merely
far from usual is down-weighted and left alone — asking would be the
over-questioning §25 warns against, and §51's 45-against-29 outlier is exactly
that case.
Nothing is ever dropped. §12 step 2: unusual data is marked for review or given
less influence, never silently deleted. A questionable interval keeps a quarter
of its recency weight, and a test asserts it is still present and still counts.
Recency decay is here too, ready for the centre in #10: newest cycle weight 1.0,
each older one 0.85 of the last.
12 new tests. The §51 acceptance cases still pass unchanged.
closes#13
Round 1 was partial and says so, pass by pass, with each not-run row carrying
why. It found two defects that 70 green unit tests could not: the primary
button crashing the app on a second tap, and two API 34 date calls on the
recalculation path that would crash every device below Android 14.
Neither is filed in the tracker. Both were found and fixed inside the batch that
introduced them, before any build left this repository, and an issue closed by
the commit that created the code would be bookkeeping rather than a record.
New standing gap: nothing has been run on a device at this project's own minSdk.
Lint is a good guard for that and is not a substitute for one run on API 26.
checkModuleBoundaries holds the dependency tables in docs/architecture/README.md
as a check: every module's permitted project dependencies, plus the rule that
domain:cycle and domain:prediction must never apply an Android plugin. core:ads
is already in the map with an empty permitted set, before the module exists —
PRODUCT_PLAN.md §34 is non-negotiable, and a guard written alongside the code it
constrains is one shaped around whatever exception somebody wanted at the time.
It lists every violation rather than the first, and refuses to report a pass
when it examined no modules at all.
THE GUARD FAILED ITS OWN FIRST PROOF
prove-guard.sh injected a forbidden dependency into :domain:prediction and the
guard reported "7 modules checked, no violations". The root project is
configured before its subprojects, so reading subprojects.configurations from
the root script saw every configuration empty — it had been green over an empty
map since the moment it was written, and would have been trusted for months.
Collection moved into afterEvaluate, and the task now throws rather than passing
if it ends up with no modules. Three proofs recorded in the architecture doc,
all re-run and all red: a domain module reaching upward, :app reaching past the
repository straight to Room, and a module with no rule being reported as
unmeasured rather than assumed fine.
TWO REAL BUGS FROM WIRING IT INTO `check`
Running the whole check for the first time turned up Android lint errors that
would have shipped:
NewApi: java.time.LocalDate#ofInstant requires API 34 (minSdk is 26)
NewApi: java.time.LocalDate#EPOCH requires API 34 (minSdk is 26)
Both are on the recalculation path. On any device below Android 14 — most of
the install base this app targets — that is a crash. Neither the unit tests nor
the API 36 emulator could see it; lint is the only thing that could.
Replaced with atZone().toLocalDate() and ofEpochDay(0), which are API 26.
Also cleared the lint warnings that were real: a redundant activity label, and
a round launcher icon declared but never referenced. The two that remain are
deliberate and now say so where the warning is read — targetSdk 36 is Play's
floor and raising it opts into untested runtime behaviour, and the -v26 mipmap
qualifier stays because removing it makes AAPT fail to resolve the icon at all.
./gradlew check now passes with 0 lint errors across all seven modules.
70 unit tests, all passing.
closes#7
The Batch 01 vertical slice from PRODUCT_PLAN.md §58 now runs on a device:
launch, log a period, it is stored, the forecast recalculates, edit or delete it
and the forecast moves again. Hilt wiring, a TodayViewModel exposing one
immutable state, and a working surface that says "Batch 01 · working surface" at
the top so nobody mistakes it for the designed Today screen, which is Batch 03.
THE DEFECT THIS FOUND, ON A DEVICE
Tapping "Started today" twice on the same day killed the app:
FATAL EXCEPTION: main
android.database.sqlite.SQLiteConstraintException: UNIQUE constraint failed:
period_records.startDate
Not a hypothetical — the crash was reproduced on emulator-5580, the fix
applied, and the same two taps then produced "That day is already logged." with
the process still alive and zero FATAL lines in logcat.
The constraint is right: a duplicate must not overwrite the original row and
lose its createdAt and source. The API around it was wrong. Repeating a tap
when you are not sure the first one registered is an ordinary thing for a person
to do, not a fault, and it must never be an exception. So the period writes
return PeriodWriteResult — Added, AlreadyRecorded, Updated, Conflict, NotFound —
and only genuine faults still throw.
editPeriod had the same hole: moving a record onto a date another record holds.
That is refused rather than merged, because merging would delete a period the
user entered and only they can settle it.
The ViewModel now installs a CoroutineExceptionHandler as a backstop. In a
health app a crash mid-write is adjacent to losing what was just entered, and a
message somebody can read beats a process that vanished. The message carries the
exception type and never a record's contents (§45).
Four regression tests pin all of it, plus two instrumented tests on a real
file-backed database that close and reopen it — what a force-stop actually does,
and something an in-memory database cannot fail.
70 unit tests and 2 instrumented tests, all passing. Release APK 1.2 MB.
closes#6
core/data is the seam between storage and everything else. Reads return domain
types, cycles are derived rather than stored, and the forecast is a function of
the data instead of a field somebody has to remember to refresh — so §11's
"recalculate after a confirmed start, after an edit, after a Not yet" is
automatic rather than three call sites.
Confirming a period is four writes in one transaction, because a partial result
is a corrupt history rather than a failed action: write the record, score the
forecast that was standing, clear the "not yet" observations it resolved, and
snapshot a fresh forecast.
THE DEFECT THIS FOUND
A test expecting one scored prediction found three. The cause was not the test:
every historical period entered during onboarding was scoring the current
forecast against a date in the past, inventing an error for a prediction nobody
had ever been shown. §16's "your predictions are getting better" would have been
populated with figures the app made up about itself — plausible ones, which is
what makes it expensive to notice.
Two rules now, both pinned by tests:
- exactly one unscored snapshot exists at a time. A forecast superseded before
its outcome was known is not a wrong forecast, and counting it lets one
cycle contribute several errors.
- a confirmed start only scores a forecast made on or before it. Anything
earlier is backfill and leaves the standing forecast alone.
Accuracy also stays quiet below three scored predictions. One lucky forecast
reading "average error: 0 days" is an overstatement, not a measurement.
THE ROOM BOUNDARY, HELD THREE WAYS
implementation rather than api on core:database; CycleRepository's constructor
internal because it names a PeriodDatabase; reads mapped to domain types in
Mappers.kt. Callers use CycleData.repository(context) and never learn Room
exists. Verified rather than asserted: grep -rn "androidx.room" app/src domain
is empty, and Room appears zero times in :app's debugCompileClasspath.
No fallbackToDestructiveMigration: it turns a forgotten migration into a silent
wipe of the user's entire cycle history on update.
Also fixed: `domain/*` inside a KDoc silently opened a nested block comment —
Kotlin block comments nest — which broke compilation in a way the error message
pointed nowhere near.
58 tests across the project, all passing.
closes#5
core/datastore holds the settings from PRODUCT_PLAN.md §10 — notification
privacy, reminder time, the three reminder toggles, biometric lock, theme, the
ads entitlement and whether onboarding finished.
Two defaults are decisions, and each has a test whose job is to stop it being
changed by accident:
- notification privacy defaults to DISCREET (§28). A default of DIRECT would
put menstrual detail on a lock screen before the user has been asked a
single question, and a notification read over a shoulder is the likeliest
real privacy breach in this product.
- fertility reminders default to off. Most users are not tracking fertility
and an unrequested ovulation notification is an unpleasant surprise.
An unrecognised stored value falls back to the SAFE option rather than to
whatever enum entry happens to be first — a rollback or a hand-edited file must
not be able to turn DISCREET into DIRECT. Tested for privacy mode, theme and an
out-of-range reminder time.
This is a separate store rather than two more Room tables, and the reason is a
deletion semantic: Delete My Data removes the health history and must leave the
settings alone. Handing a user back a weaker privacy setting at the exact
moment they are exercising a privacy control is the worst possible time to do
it, and separate stores make the correct behaviour the easy one.
The repository takes a DataStore rather than a Context, so its 10 tests run on
the JVM against a temporary file — no emulator, no Robolectric. The Android
instance is supplied by DI at the app layer, the only place that should know
where a file lives.
41 tests across the project, all passing.
closes#4
core/database holds the four entities from PRODUCT_PLAN.md §10 —
period_records, spotting_records, prediction_records, not_yet_observations —
with DAOs returning Flow, epoch-day/epoch-milli converters, and the schema
exported to core/database/schemas and committed.
Three constraints are structural rather than remembered:
- startDate is UNIQUE and inserts ABORT rather than REPLACE. REPLACE would
delete the original row with its createdAt and source; §14 says health
history is never modified silently.
- spotting has its own table, so no query for periods can reach it. §25: it
must never start or reset a cycle.
- a prediction snapshot can be scored but not rewritten — score() sets only
actualStartDate and absoluteErrorDays. A snapshot editable after the fact
can only ever report that the app was right, which would make §16's whole
accuracy feature a lie.
deleteEverything() is one transaction and the only bulk delete in the module: a
partial wipe leaves the cycle reconstructible from the tables the user asked to
be rid of.
14 tests, on the JVM under Robolectric — no emulator.
THE SCHEMA GUARD, AND WHY IT IS A SCRIPT
SchemaTest was written as a drift guard and proved not to be one. Room
regenerates the schema export during compilation, so adding a column to
PeriodRecordEntity without bumping VERSION leaves the suite green while the
committed schema quietly changes underneath it. That was not reasoned about, it
was run: the column was added, 1.json gained it, and every test passed. On a
device that is "Room cannot verify the data integrity" — a crash on update,
after shipping.
scripts/schema-guard.sh asks git instead, which Room cannot overwrite. Proved
both ways before being trusted: green on a clean tree, exit 1 on the injected
drift. It runs in pre-commit when an entity or the schema directory is staged,
and the hook treats its exit 2 as a refusal.
SchemaTest keeps its four tests and now documents what it does not catch.
Room's own MigrationTestHelper is not used: every constructor needs an
Instrumentation and schema assets, and AGP 9's library source-set DSL throws
DefaultAndroidLibrarySourceSet_Decorated cannot be cast to
AndroidLibrarySourceSet when you add an asset directory. Recorded so the next
person does not spend the afternoon on it.
Docs updated in this commit, as their triggers required: the migration table
now has its version 1 row and the trap that makes such tables go stale, TOOLS
explains the seventh script, and the hooks README lists the new guard.
closes#3
doc-claims.sh reported 19 claimed paths that do not exist. Every one was a
deliberate forward reference — the planned modules in the architecture table,
the two documents the trust map records as absent on purpose, and the release
script Period declined.
A backticked path is read as a claim the file is there, so a document saying
"core/database does not exist yet" was asserting the opposite of what it meant.
docs/history/BATCH_LEDGER.md already records the idiom for this case; it is now
applied and stated where it is used, so the next forward reference does not
reintroduce the failure.
Also corrects WORK_CYCLE.md, inherited from the template, which pointed at a
scripts/release.sh this project does not have. A release here is a signed AAB
and a Play submission, so the security checklist carries that procedure.
doc-claims.sh now reports 152 claimed paths, all present, across 20 files.
Period was a bare directory holding one 2,527-line specification, with no git
repository, no tracker and no documentation convention. This is the adoption
from Projects/Template/START-HERE-New-Project.md, plus a project that compiles
so the hooks and future guards have something real to run against.
Documents. scaffold.sh created 19 paths, 0 skipped. The specification moved to
docs/planning/PRODUCT_PLAN.md unchanged in substance, with a status header; the
capitalised Docs/ is gone. Every scaffolded document was filled in for Period.
docs/OPERATIONS.md deleted — an offline app is not a deployed service.
DOC_TRUST_MAP.md written last, describing what is actually here, including what
this project deliberately does not have.
Code. Four Gradle modules. domain/cycle and domain/prediction are kotlin("jvm")
and cannot see the Android SDK, so the engine is testable without an emulator —
17 tests pass, 12 of them the acceptance cases from PRODUCT_PLAN.md §51.
BaselinePredictionEngine is a robust-median prototype and explicitly not the
product; it exists so Batch 02's replacement can be shown to be better rather
than merely different.
Versions verified against their official sources today rather than inherited
from the specification's own numbers, which that document asks for: Kotlin
2.4.10, AGP 9.3.1, Gradle 9.7.0, Compose BOM 2026.08.00, Room 2.8.4, Hilt
2.60.1. AGP 9 ships Kotlin built in, so org.jetbrains.kotlin.android is no
longer applied. compileSdk is 37 because current AndroidX requires it; targetSdk
stays 36, Play's floor from 2026-08-31, and the difference is deliberate.
Six scripts taken into scripts/; the rest declined and named in docs/TOOLS.md.
Three hooks in .githooks/, with pre-commit adapted to Gradle.
closes#1closes#2
null