Answering the August 28 review

You set the test yourself: if you were buying an icon set of fifty icons, you could imagine all of these being in that package, as though one mind drew them. The app now passes it.

Phosphor is the single icon library. Its outlines are built into the app as drawing instructions rather than pasted in as pictures, so nothing in the set can drift away from the rest of it. Weight is used for one job only: the tab bar fills the tab you are on, and every icon inside a screen is drawn at the same line weight. The candidate pool was cut from 152 icons to 60, and then each recurring thing was given exactly one icon, everywhere it appears. A reading circle is the seated ring. A one-to-one is two people. Feedback is an open envelope. There is one microphone in the product.

Seven of your other notes fold into this one pass. The generic bell is the library's bell. The four live-session types read as one family, differentiated by their own mark rather than by line weight. The lesson card carries a video mark and a running-time mark, so you can tell it is a lesson without reading the heading. Kamran and Yasmine settled the six picks that were genuinely open on a shared board before any of it was drawn.

The five states already had names — awaiting review, in review, feedback ready, returned, taken back. Each now has its own icon on the row, drawn from the same library as everything else: an hourglass, headphones, an open envelope, a u-turn arrow, an open hand.

The status wording stays as text in the pill beside it. The icon does not replace the words, it gives you something to scan for before you read them. The detailed pen-and-ink mark that made "returned" unreadable — the one drawing with far more detail than anything around it — is gone.

The floral register is retired and its eighteen painted image files are deleted from the build. What replaces it is not a pattern printed behind the numbers — it is a panel the student fills.

The fifteen weeks are now one octagon-and-square tessellation with a seat for each week. A finished week is a plum stone set into its seat, the week you are in is the one gold stone, and the small square tiles between the seats gild in as their neighbours complete. The numerals stay live text over the geometry, so the grid is still readable as weeks.

The level mark is the QuranFlow logo's arch carrying the level numeral inside it. It is used only where one specific numbered level is the subject — placement, and the screens that announce a level — and it is never shown empty, because empty at that size it reads as a zero. The traced arch is a placeholder for the brand team's own vector.

Today's corner ornament was cut rather than redrawn: it only ever appeared on the feedback card in the week feedback had just landed, which is not one of the states captured below, and no ornament replaces it.

Green means completed. That is the whole rule now, and nothing else uses it.

The collision was that reading circles were wearing green as their identity, so green meant "a circle" in one place and "done" in another. The reading circle takes lapis blue instead — its tile wash, its disc, its time tag and its dot in the calendar legend. Live and starting-now went plum, as did an active enrollment. A closed week went muted. After the sweep, exactly one green survives anywhere in the app: the word "Completed" in your learning history, which is the job green keeps. The move also lifted two colour contrast pairings that were failing.

Arabic always ends on white. A passage card now starts on pure white and settles into warm paper at its foot, instead of finishing on a lilac tint. Every passage surface on the record-and-feedback loop was checked in the browser. One exception is on the record: the Lesson screen draws its own passage card and is shared with the older build, so it was left alone and needs its own ruling.

The dawn wash now runs at one muted strength on every root screen, the same quiet one the You, Sessions and Learn screens were already using, rather than shifting with the state of the week.

They stay, everywhere. Gold is the action colour on every button in the app, and hierarchy is expressed by fill against outline — a filled gold button is the act, an outlined one is the alternative — never by a second colour.

Worth recording, because it was built and then withdrawn on the same day: a pass that demoted secondary buttons to a plum fill created two competing filled button colours on one screen, which is the pattern both Apple's tint colour and Material's emphasis levels exist to avoid. It was reverted. The v15 gold buttons you liked are the ones shipping.

Nothing changed here, so there is nothing to show.

Twelve items from your list, each one its own small fix. Your words on each are quoted with it.

The grade. A 1 to 5 grade in half steps now sits at the send seam of the review, directly above Send. Half steps rather than tenths, because as Kamran put it, if there is no difference in meaning between 3.1 and 3.2 there is no point in offering it. It is labelled internal on the screen and the student does not see it. This is deliberately separate from the Academic Lead's rubric proposal, which stays in its own exploration.

Her hours needed nothing built. The screen already models what you described: submission length plus feedback length plus a thinking margin, the scheduled classes added on, and the adjust-hours control with a reason. The calculation engine, the reassignment that credits a substitute, and the monthly report that becomes the invoice are backend work and go to the admin track, not this round. The only thing that moved on that screen is colour.

Curriculum access is likewise validated and unchanged. Every coach sees the whole library, both years and every level, with the same lesson view the student has. Nothing was asked for and nothing was changed, so there is no pair here.

The entrance you could not find is the coach entrance, and the hub landing page carries a coach card of its own this time, both as the app and as a guided tour. Last round it was missing because the preview was not a full deploy.

Her side also takes the same icon and colour work as everything else: the seated ring, two people, the lapis reading circle, plum for live.

You granted Kamran's frame and held your preference, so the fix took the part that costs nothing. The Start recording card moved to sit directly under the passage, and the clarifications — the due date, the take-back window, what your last review flagged — moved below it. The assignment and its action are now paired, without giving up the card grammar or pretending the fold can be controlled across devices. Measured on an 844-pixel-tall screen, the button moved from 881 pixels down the page to 518.

A separate finding from a walk of the whole flow: the microphone permission screen — the one you said you really liked, where the white card stands against the shaded background — was finished but orphaned. Nothing in the app opened it. It is now the first run of the record flow: the first time you record, it asks, and granting replaces it with the recorder rather than leaving it behind you. It draws the recorder's own microphone, which is the consistency you asked for.

The profile checklist. You said the old one was cleaner because you can enumerate what remains, and the new one reads like an ID card, but that either works. We left it. Your preference is on file, so if that screen is reworked for another reason it starts from the version you preferred.

The default avatar. You said the app is female-coded and then the profile icon is a man. The man you saw was the old icon set, which is retired. The person icon now used on the You tab and in every person seat is the library's generic user mark, which is not gendered — no beard, no head covering, no shoulders that read either way. Drawn faceless avatars were explored and withdrawn: a designed hijabi or bearded figure makes a choice about a student on her behalf, on a screen where the app knows nothing about her. Kamran ruled: no custom avatars.

Gold. Restated once, because it was the other thing you asked for and it is easy to read a non-change as an oversight: gold buttons stay everywhere, and hierarchy is fill against outline rather than a second colour.

Two items are decided but not built, and both wait on the same thing.

Assignment naming. A week mixes ayat across surahs to drill one rule, so the surah-and-verse titles in the current sample content are wrong. The student-facing pattern is "Week 8 assignment", with the admin dictating every display title from the backend as free text. Kamran's note for the developers is recorded: these are backend content fields on a consistent schema — title, author, running time — uploaded structurally, not app copy and not typed in one at a time.

The shape of a lesson. A week's lesson is one titled unit holding up to about three titled videos: the concept video, the walkthrough for that week's submission, and a practice-with-me or extra-explanation video. The current sample content does not have that shape.

The one thing blocking it: the level-lessons and curriculum document, which is the document that actually gets entered into the backend. Both changes are a pass over the sample content on every screen that names an assignment or a lesson, and neither can start until that document exists. Nothing else in this round is waiting on anything.