Accuracy and fidelity
How DocViewKit distinguishes parser success, visible approximation, and native visual acceptance.
Three different claims
| Level | What it proves | What it does not prove |
|---|---|---|
| Parser and protocol | The input produces bounded document objects and source references | Native visual fidelity |
| Browser render | The public Viewer draws usable content and interactions | Pixel identity with an Office application |
| Native visual acceptance | Reviewed pages or local regions match a named native application under a recorded environment | Universal fidelity for every producer and feature |
Open is not a fidelity result
A document opening successfully is not proof that every authored object, style, page break, formula, animation, or embedded resource was rendered.
Best-effort rendering stays observable
An isolated unknown or incompatible resource should not hide otherwise usable document content. Supported fallbacks remain source-mapped and emit approximate or omitted diagnostics; structural corruption, unsafe expansion, or a document with no meaningful output may still fail the open.