>The sequence 1) table, 2) view, 3) AGAIN NOUPDATE, does not seem to produce the probmem.
>
>The sequence we actually had last week, 1) table, 2) AGAIN NOUPDATE, and finally 3) the view does produce the problem.

Ah, so it looks like VFP is remembering the NOUPDATE from the most recent use of the table and applying it to the View, even though the base table is read-write.

That's plausible in terms of how a feaure (OK perhaps it's a bug) in VFP might be causing the problem.