On Tue, Apr 2, 2013 at 2:01 PM, Samantha Atkins <sjatk...@me.com> wrote: > Natural keys are in user data space. Thus they are not guaranteed invariant > and therefore cannot serve as persistent identity.
That is true, but irrelevant in most real world cases. Also, nothing is keeping you from using an extra marker if/when you need to provide an invariant lookup. > Also natural keys have the considerable defect of being of different type and > arity per logical entity type. This means that very nice things like > dynamic relationships and persistent collections of references to multiple > types of things cannot be built. It also increases ORM and data cache > complexity. OO evangelism. > These are considerable weaknesses. > > You cannot guess a uuid from say uuid.uuid4(). If you think you can I look > forward to seeing your proof. I never claimed that. I said that postgresql random() can be guessed, which it can, since it's based on lrand48. merlin -- Sent via pgsql-general mailing list (pgsql-general@postgresql.org) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-general
