Re: Much Ado About COUNT(*)

Jeff Davis <jdavis-pgsql(at)empires(dot)org> writes:
> I almost think to not supply an MVCC system would break the "I" in ACID,
> would it not?

Certainly not; ACID was a recognized goal long before anyone thought of
MVCC. You do need much more locking to make it work without MVCC,
though --- for instance, a reader that is interested in a just-modified
row has to block until the writer completes or rolls back.

People who hang around Postgres too long tend to think that MVCC is the
obviously correct way to do things, but much of the rest of the world
thinks differently ;-)

regards, tom lane