How can I use these cover images when the scratch labels disagree?

Sample found and numbered. Quality is checking a faint line close to the edge. The big bright streak changes location in the pictures, so I have separated those observations.

Put a status on the unresolved line. Otherwise the first person tidying the folder will decide whether it belongs in good or bad for you.

Will the physical sample keep that unresolved status too?

Yes, in a marked reference box. I have stopped using it in student exercises while quality decides. They still have to agree the viewing conditions.

Keep a wide photo with the close-up. That helps a beginner find the same edge on the real cover, though quality still needs to define how they inspect it.

And record the camera and lighting arrangement on the next set. We get 'same lighting' written on things when the lamp has plainly wandered somewhere else between sessions.

Quality accepted the faint line under the agreed appearance check. We have recorded its location rather than pretending it isn't there. I'm now comparing the accepted cover with a known rejected sample.

Who owns future label changes? A clear decision now won't stop the next engineer from casually renaming the image during a lighting trial.

Quality technician owns the reference register. Engineering can add development images but cannot replace the agreed judgement in that set. Thank you, Oliver; that ownership was missing.

Can students see why the faint line was accepted, not just that it was?

I'd include quality's explanation with the viewing conditions, Beth. Not a universal rule that faint lines pass. Otherwise your teaching example will become a shortcut applied to a different finish.

Tooling found a corner could hang on the nest rim. Revised clearance is proposed. We're keeping those inconsistent shots as development evidence and waiting for repeat presentations before choosing lighting.

Label the nest revision on those pictures too. 'Before' and 'after' become surprisingly unhelpful once there are three afters.

Done. Revised nest is R2; original R1 remains identified in the development folder. Repeat images are more consistent so far, with the rejected sample's mark still visible.

How many different covers have you tried in R2? One repeated part is useful for checking presentation, but it won't show whether other in-tolerance covers seat differently.

Only the two references so far. More samples are included in the next checks. I should say these two repeat more consistently, not that the whole setup is settled.

That's a useful limited result. Keep the wider sample check assigned before anyone treats the new pictures as approval to finish commissioning.

Thanks for the R1/R2 detail. Much easier to follow which pictures explain which problem.