Mobile · In development
The decision
Health-domain red lines are checked automatically, not proofread
No promises of outcome, no medical vocabulary, no statistics, neither inside the app nor in the store listing copy. These rules are not held up by vigilance: they are an automated check the verification gate runs, alongside the tests. The reason is simple: health copy reads very well and reads very badly. A phrase drifting towards a promise goes unnoticed on the tenth pass, especially when it is your own writing. A check does not get tired. The constraint also covers text produced by the language model, which is the likeliest source of an unintentionally clinical sentence.
The problem
Tracking symptoms day to day means an app that records, summarises and reminds. In health, each of those three is a minefield: a summary can read as a diagnosis, a reminder as an instruction, a chart as a promise of results. The hard part is not building the tool, it is building a useful one that never crosses that line.
My role
Solo: the iOS app, the server-side summarising service, in-app purchases, and the verification chain that guards all of it.
What was hard
The name. The one chosen sits ONE letter away from a registered French trademark, used in the same country, for a nearby audience. The class the app targets is free, the neighbouring mark does not cover it, and the spelling differs. But none of those three arguments is worth an IP lawyer’s opinion, and I have not taken one yet. So the name is in the app, and NOTHING is reserved: no domain, no account, no filing. It is the most interesting difficulty on the project precisely because it is not technical, and because it is still open.
Where it stands
In preparation for the App Store. The app runs on a real device; the store listing is a draft awaiting review, and the naming decision is waiting on trademark advice. Nothing is published.