Tom Brandt <[log in to unmask]> quickly replies:

> At 09:59 AM 07/17/00, Jim Phillips wrote:
> >My first thought was your second method, since we all know that the
DBDELETE
> >will fail if there are details associated with it.  I have no idea which
is
> >more expensive, but the problem with relying on something to be
consistently
> >broke is that someone may someday decide to fix it and then you're up the
> >creek without a paddle (or your successor is).  In other words, there is
no
> >guarantee that DBDELETE will continue to fail when there are associated
> >details.  HP may decide to fix that tomorrow, making the default DBDELETE
> >action be to delete both the master and associated details (not very
likely,
> >but you never know),  and then where would you be?
>
> I don't think that DBDELETE's current behavior is "broken" - I think it
> works exactly the way it is supposed to.  It is deleting an entry, not a
chain.

Oh, my.  I should have put quotes around "consistently broken".  I don't
think DBDELETE's current behavior is broken either, but I still don't like
relying on a failure to keep me safe from doing something I really don't
want to do; i.e., deleting the details and master when I want to keep them.

> If HP did change DBDELETE to delete chains associated with a master, given
> the conservative way they change IMAGE, they would no doubt use another
> mode for DBDELETE so the current mode 1 DBDELETE would work the same it
> always has.

You're probably right, but I still don't like relying on a negative, I'd
rather rely on a positive.


Jim Phillips                            Manager of Information Systems
E-Mail: [log in to unmask]     Therm-O-Link, Inc.
Phone: (330) 527-2124                   P. O. Box 285
  Fax: (330) 527-2123                   10513 Freedom Street
  Web: http://www.tolwire.com  Garrettsville, Ohio  44231