Everything I did not touch

Simple 2.4 went live on the App Store on 5 October. The headline is the AI Assistant: it now plans, breaks tasks down into small steps with time estimates, and starts a focus session with you. By the start of October that work was done, and the release looked ready.

But a release is more than its headline.

The audit

A big feature gets read line by line and tested on device. The screens around it do not. They worked last month, so they are assumed to work this month. In a planner, those screens are most of the app: onboarding, the paywall, the task list, the timeline, the widgets, focus, calendar sync.

So before 2.4 went to review, the whole app had an audit with one rule for its scope: everything the 2.4 work did not touch. Five passes, one for each area, every finding read in the code. It came back with thirty-two items, sorted by how much harm they could do.

What it found

Some of it was the dull kind of debt every app collects. More than a thousand print calls shipping in the release build. Hundreds of duplicate entries across the sixteen string catalogues. Buttons with no label for VoiceOver.

Some of it was not dull at all.

The "Clear completed" button deleted every finished task it could find, not just the ones in the list you were looking at, and it did not ask first. If the local database failed to load, for example after a failed migration, the app deleted the database file and started again. Editing a finished task quietly marked it as unfinished. Editing one occurrence of a repeating activity could duplicate the whole series. The paywall could offer a free trial to someone who was no longer eligible for one.

None of these were new. They had sat in screens I used every day, in code paths I did not happen to walk. That is the uncomfortable part of building a planner for my own brain: being its most loyal user makes me a poor tester of it. I walk the same paths through it, and the bugs live on the ones I do not.

How it got fixed

The findings went in a fixed order. Safety first: the things that lose data or break App Review rules. Then money, then the first-run flow, then daily use, then hygiene.

The database wipe became three retries and then moving the unreadable file aside, never deleting it. "Clear completed" now clears only the open list and asks before it does. Repeating activities got a proper series identity, so changing one occurrence changes one occurrence. The trial offer is checked live against the account, and "Try it free" only appears when there is a real trial behind it.

Then the parts people feel every day. Overdue tasks used to vanish from Today, which then cheerfully said you were on top of everything. Now they sit at the top with a button to move them all to today. The timeline was reloading its data from the database every second on the main thread; it now listens for changes and refreshes once a minute. Onboarding asked for an account before showing anything useful, and returning users who signed in with Apple were sent through the whole thing again. The account now comes after the quiz and the AI demo, and a returning user goes straight to their day.

Widgets had their own list. Ticking a task off on the Home Screen did not save until the app was opened, so a reminder could still fire for something already done. The widget now writes to the shared database directly.

By 4 October every item on the list was ticked off, and build 52 was ready for review. A day later 2.4 was on the App Store, with a "Fixed" section in the release notes that lists several of them.

What I keep from it

The lesson is not that audits are good. It is narrower than that. A release draws attention to what changed, and bugs collect in what did not. Thirty-two of them, in this case, in an app I open every day.

If you plan your days with Simple, 2.4 is the version where the boring parts got the same care as the new ones. What the app is for, and how it fits an ADHD day, is laid out in the guide to what actually works in an ADHD planner app.

Simple ADHD Planner is on the App Store.