> The short answer is that Adam's statement is wrong, or at least
> misleading.
Sorry Tom, I wasn't trying to do either. Joshua Drake (who I understand
to be a reliable source of postgresql information) said that applying
a sequence to a column after creation created issues, versus using the
serial type which did not. That seemed misleading to me, since it's
*exactly* what pg_dump does in 8.3. All I did was point that out,
which I'd hardly call "complaining" and definitely not wrong.
My point was that there was nothing special about serial in 8.3
Nothing "misleading" about that either.