Does anybody have a definitive Cause and Corrective Action for a problem that
seems to occue infrequently with oxiding?  

Specifically a detail with a power or ground plane and a signal layer
opposite will oxide fine on the power/ground side, but will occasionally
oxide incompletely on the other side.

I've always thought it was basically a cleaning problem, and I know it has
occurred on a variety of coppers and substrates.  But on occasion it has been
very persistent, unable to be improved by strip and re-oxide for example.

Very frustrating to the PWB producer and both the laminator and the chemical
supplier get accused of having something wrong with their products.

Any ideas?