Product · Scope
The features we deliberately did not build
Every product page tells you what the software does. This is the other list: real requests we evaluated, priced, and refused, with the reasoning left in the decision docs where you can read it.
Supplier entities: the point of no return
InventoryCheck's spec has a hard fence: no purchase orders, no supplier management, no valuation, no multi-currency. When the restock-procurement roadmap pushed against it, we decomposed the question into levels instead of arguing about a vibe. A read-only estimate report, "what's on the shelf worth, roughly, at last-known prices", passed the fence, because a labeled estimate is not an accounting ledger. Order sheets, supplier entities, and approval workflows did not.
The rule we wrote down: supplier entities are treated as the point of no return, and they're never taken "on a hypothetical." The moment an inventory app stores suppliers, users reasonably expect purchasing, then invoicing, then a ledger, and you've built half an ERP badly instead of one inventory tool well. The fence isn't a feature decision; it's the thing that keeps feature decisions cheap.
Why "no" gets written down
A small team can't win on breadth, so scope discipline is load-bearing architecture, not a marketing posture. Each refusal lands in a decision document with the date, the alternatives weighed, and the level at which it would reopen, usually "when a paying customer demonstrates the demand." That's not stubbornness; it's how fourteen apps stay maintainable by three people. A feature we can't maintain is a liability sold at license price.
The same fence logic shows up across the suite. DutyCheck's on-call module was scoped to actual standby duty, not a general availability tracker, because the neighbouring features each pull in their own compliance surface. ArbeitszeitCheck exports DATEV files but will never transmit to DATEV directly: the pipe-separated CSV your payroll office already knows how to import beats a certification project we can't staff.
What this means if you're evaluating us
Read the fences as a feature. A vendor that documents its boundaries gives you a product that is the same shape in three years, the thing you bought, maintained, rather than the thing it accidentally became. And when a fence is genuinely in your way, the paths are also documented: the code is AGPL, the exports are open formats, and commissioned features exist for needs we judge worth maintaining.
The decision docs referenced here are internal working files, not marketing artifacts, which is exactly why they make good reading if you want to know how the product will behave when your request lands somewhere near a boundary.