The gate at the exit — how a green build still ships a wrong page
What was tried
A pipeline simulation with a document-unlock step that asks the reader to judge, from the published page alone, whether that run is wrong — before revealing what happened underneath. Three switches (view, unlock password, exit gate) give four comparable states, and each number can be opened to walk its provenance back to the source documents.
What a reader sees
last updatedpipeline: success5 sources
Is this page wrong?This is everything a reader sees — 4 numbers, a date, and the word "success". Decide before you press "Inside the pipeline".
What was learned
What this round actually established is about the design, not about readers. The first version showed all three zones at once and killed its own point: the source zone turned red for unreadable files and the pipeline zone announced how many were skipped, when the whole claim is that nothing tells you. A real reader only ever sees the published page, so that view has to be the default and the four-state answer table has to start hidden. Second: if the gate halts the pipeline then no new page shipped, so showing a page full of holes next to the words "pipeline halted" is incoherent.
What fell short
Nobody but the author has tried it. The assumption the entire entry hangs on — that a person cannot tell from the published page alone — has not been tested on a single reader, so it remains the designer's belief rather than a result. It is also a simulation with precomputed outcomes rather than a real pipeline run, so it says nothing about whether such a gate catches every real case. And the gate guarantees only that a number has a source, never that the source is correct: if the source document itself is wrong, the gate sees nothing.