Filtering and identifiers and logging.

[Rod Widdowson]

> >Ah, OK, I think I misread Rod to have been saying that he had verified
> >the opposite.
> 
> I also thought so, so I tested it. He was looking at the code, but never
> tried it. So that's "better" in a sense.

Both actually.  I just ported the code over to V3 and tested it there (where
the schema is not validated)

R