"Greg Sabino Mullane" <greg@turnstep.com> writes:
> That particular example is a very poor one for illustrating your
> point. You severely underestimate "get away with" for the implicit
> cast changes in 8.3. This was a really big deal for many, many users
> of Postgres, and continues to cause many problems to this day.
Actually, that was *exactly* my point. What's being blithely discussed
here is to make tens of changes that would be equally as painful as that
one, all at once. You really think that would fly?
regards, tom lane