Like I was saying to Mike, it's simply a guard against performance
issues. It's like omitting inverse relationships
when you know that it would be suicide if anyone tried to traverse
(or inadvertently tripped) them.
e.g. transaction -> user_agent is a fine relationship to model but
would you even dream of modeling the inverse relationship?
Maybe you would if you only made one sale a week.
I'm with you here, but the same would be true of the non-flattened,
right? If you don't want to fault the relationship, don't model it,
but that seems to me more of a pragmatic decision about the existence
(or not) of the relationship rather than specially related to many-to-
many flattening. As far as I know, the faulting rules should be the
same.
ms
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Webobjects-dev mailing list ([email protected])
Help/Unsubscribe/Update your Subscription:
http://lists.apple.com/mailman/options/webobjects-dev/archive%40mail-archive.com
This email sent to [EMAIL PROTECTED]
