Re: The Practical Benefits of the Relational Model

"Paul G. Brown" <paul_geoffrey_brown_at_yahoo.com> wrote in message news:57da7b56.0211032027.7e218417_at_posting.google.com...
> "Paul Vernon" <paul.vernon_at_ukk.ibmm.comm> wrote in message
news:<aq3r28$k8a$3_at_sp15at20.hursley.ibm.com>...
> > Come on Paul, keep with the program. ;-)
>
> It's been a big program.

True. Maybe I'll summerise it all in a new thread..., or maybe not

> But (lucky for me) implementing 'multi-assignment operators' will still
> require all of the considerable body of theory (dependency graphs) and
> engineering practice (locking structures, logging, recovery, blah blah
blah)
> I've been working with.

Lucky you. But yes, you will still need all that (well, at least a lot of it I guess) under the covers. In fact, I would strongly suggest a developer level interface where external triggered actions can occur due to relvar value changes. E.g. when I INSERT into my INSTANT_MESSAGE relvar, an message is sent, or if the message fails, then the insert would also fail. Such stuff would indeed need transactions, just not user visible ones.

> I dunno. I guess I don't see a logical difference here.
>
> What can be done with 'multiple assignment operators' that cannot be
> done with explicit transaction boundaries? (Or should I just get with the
> program a bit more?)

I'll refer you to Bob's answer.

Regards
Paul Vernon
Business Intelligence, IBM Global Services Received on Mon Nov 04 2002 - 18:52:55 CET