Re: OO and relation "impedance mismatch"

Kenneth Downs wrote:
> > I look forward to it. As I have said before, I know for a fact
that
> > your approach wouldn't work 100% for Oracle (which doesn't even
support
> > dirty reads), but I can't be so sure about other DBMSs.
>
> I have not used Oracle yet seriously, but considering most colleagues
who
> have tell me it can "everything", I would be willing to be a dollar
that it
> can somehow do it by another name. For instance, DB/2 calls it a
> "non-repeatable read", which name is self-explanatory, though perhaps
it
> should have been called "potentially non-repeatable read."

You'd lose that dollar. Oracle doesn't do dirty reads at all:

http://download-west.oracle.com/docs/cd/B12037_01/server.101/b10743/consist.htm#sthref1924

This may be why to me, as an Oracle junkie, your scheme to dispense with all declarative constraints in favour of procedural triggers sounds rather crazy. It hasn't escaped my attention that after several days of this exchange, no one else has backed me up on this!

However, if your goal is platform independence, and if you want Oracle to be one of the possible platforms, then the lack of dirty reads may cause you some problems I guess. Received on Thu Oct 07 2004 - 04:17:07 CDT