Lexicon
The words we use for the gates between a working prototype and a launched product. Each one is defined once, here, and means the same thing everywhere else on this site.
Two of these are ours. The rest are phrases already in circulation that nobody had pinned to a definition — so we did.
- Launch gate
Any check that stands between "it works on my machine" and real users — App Store review, security audit, or Australian regulation.
The word is doing real work: a gate is binary and someone else operates it. No amount of polish gets you through one, and no amount of AI assistance changes who decides. Naming them as a set is what lets a founder plan for them instead of meeting them one surprise at a time.
- Two-bucket report
An audit finding is either a launch blocker or an after-launch item; there is no third bucket and no severity score.
Five-point severity scales let everyone agree a finding matters and disagree about whether to act. Two buckets force the only decision a founder actually has to make this week, and make the report legible to someone who is not technical.
- Classification creepour coinage
A feature added in good faith — a dose calculator, a "share with your doctor" button — that silently moves an app into medical-device territory under the intended-use rule.
There is no code smell, no warning, and no build failure. The app's legal category changes because of what it now claims to help someone do, and the team usually finds out during procurement or a funding round rather than during development. It is the most expensive thing on this list and, until now, the only one with no name.
The page explaining this one is still to be written.
- Wrapper taxour coinage
The cost of paying to wrap a web app, failing Guideline 4.2, and then paying again for the native layer that would have passed the first time.
It is a tax because it is avoidable and because it is paid twice. The decision that incurs it is made early and cheaply — wrap the existing site — and the bill arrives weeks later at the review queue, by which point the wrap has been paid for and cannot be salvaged.
- The September problem
The annual iOS and Android release that breaks apps nobody has maintained since the last one.
Every September, Apple ships a major iOS version and the APIs an app was built against shift underneath it. An app with no maintenance rhythm does not decay gradually; it works until a date it was never told about, and then it does not.
The page explaining this one is still to be written.
- The twelve-tester wall
Google Play's closed-testing requirement — twelve testers running the app for fourteen consecutive days — which blocks first-time Android publishers who do not yet have an audience.
It is a wall rather than a hurdle because no amount of engineering gets past it. Google evaluates genuine engagement, not sign-ups, so the requirement is structural: you need users before you are allowed users.
The page explaining this one is still to be written.