Re: Opinions on INSTEAD OF TRIGGERS(and probably also "old triggers)

Hi Frank:

    I've only used these in a limited way. I wouldn't be happy to have developers writing them. I create views on multiple tables and let the developers use the views. Then I write the instead of triggers on the views myself. The developers don't see what's going on. They just have a view which is simple and quick.

    I'm in the same boat you're in. I'd like to here others' first hand experiences with these. As it turns out there are a couple I have to do tomorrow, so maybe I'll learn more then.

Van

"Frank" <franjoe_at_frisurf.no> wrote in message news:DU1t6.1877$mh4.83602_at_news3.oke.nextra.no...
> Hi!
>
> I haven't tried these new triggers yet and would like opinions/experiences
> on their use.
>
> I'm slightly worried that they hide logic in a unfortunate way. I feel
 that
> the old triggers do, of course sometimes it is ok to use them (like
 keeping
> denormalized information up to date in another table). But I feel that
 their
> actions are hidden from the developers; you do a Insert, but there are
 more
> logic going on than you see, without digging all_triggers.
> To me these new triggers have the same tendencies. One situation that I
 see
> use of them is if you have a legacy Oracle system, and you must implement
> views over a partially redesigned datamodel, and in order to keep old
> insert/update statements running you choose to implement instead of
> triggers. But is it a good idea to start using them alot in new
> applications?
>
>
> Also in the old triggers there were complications (read: mutating :-) when
> trying to implement too much business logic in the trigger, it is possible
> to work around it, but the developer must keep those problems in mind.
>
> Have anyone any "best practise" to choose to use instead of triggers for,
> compared to eg. using plsql procedures to insert/update? Are they
 necessary
> for OO databases for instance?
>
> Frank
>
>
>
>
>
>
Received on Mon Mar 19 2001 - 19:07:36 CST