The Day a Rejection Page Showed No Score

Every rejected product on SleekNova has a lab note explaining why it did not make it. Two things on those pages were wrong, and this is what we changed.

The score stopped showing. Yesterday we built a proper field to hold the score a rejected product earned, and a band on the page to show it against our bar of 8.0. The field was empty on all 198 records the day it shipped. So for one day, a rejection page showed no score at all, where 88 of them had previously stated one inside a sentence. We said so at the time and this closes it: 177 of the 198 now carry the score their own note proves, and it renders in the band with the shortfall beside it.

Some records could not be published as we hold them. The subject line of a lab note is meant to name the thing we tested. On the rejected notes it had drifted into carrying the verdict instead, and 21 of them carried a long dash, which our house style bans on any page a reader sees. Two carried a subject so long our own importer would have refused the record, which meant those two could not be repaired by re-import at all.

Every subject line now names the product and its maker, and nothing else. The finding it used to carry lives in the sentence directly beneath it, where the note already made the same point.

Some records still show no score, on purpose. A score is only published where the note itself proves it: the figure has to name itself as the final weighted score, and it has to sit below our 8.0 bar. Where a note records only a trust reading, or states a figure it never reached, the field stays empty and the page reads complete without it. Seventeen of those twenty one were rejected on our trust floor rather than on the total, so a total was never the reason and was often never written down.

We checked the numbers two ways that do not depend on each other. Eighty three of the notes print their own arithmetic; recomputing all eighty three from the weights our scoring code holds gave the same answer as reading the sentence, on eighty three of eighty three. No score was rounded, inferred or reconstructed. Where a note published a rounded figure and its exact total, we kept the one the note itself published.

No sentence in any of the 198 notes was rewritten. Every word on every page is the word that was already there.