On Fri, 2002-08-02 at 08:55, Curt Sampson wrote:
> On Fri, 2 Aug 2002, Marc G. Fournier wrote:
>
> > Isn't inheritance kinda one of those things that is required in order to
> > be consider ourselves ORBDMS, which we do classify our selves as being?
>
> Well, it depends on what you call an ORDBMS. By the standards of
> Date and Darwen in _The Third Manifesto_,
Is _The Third Manifesto_ available online ?
> table inheritance is not
> required and is in fact discouraged as a feature trivially implemented
> with views, foreign keys and constraints. (Though that does not
> mean that posgresql currently has an implementation of these that
> will make it trivial.)
Could you brief me why do they discourage a syntactical frontent to a
feature that is trivially implemented ?
If it is just views. foreign keys and constraints anyway, it should not
add compexity to implementation.
OTOH, stating explicitly what you mean, can give the system extra hints
for making good optimisation decisions.
-------------
Hannu