Sorry if this is a repeat topic. I looked around and couldn't see it anywhere else.

Here's my question: Quite often, I hear other developers saying how they've set up various relationships because they really wanted to "avoid performing a find". Why is that?

Unless a relationship is already set up for structure purposes, it is so much easier to set up a find in a script, rather than add to an already overpopulated relationship graph. Is there something performance-related, or is it an attempt to create queries like you might find in other databases?

Well there are the obvious situations...

1) Performing a find requires the user to trigger the script by pushing a button, which is somewhat less efficient, and more prone to errors if failing to do so results in the erroneous information being displayed.

2) Setting up a relationship, e.g. in a portal, doesn't alter the current found set, which is sometimes desirable.

3) Relationships can help quickly navigate between related data in different tables, & populate value lists.

Yes... but relationships are useless when the found set is based on a varying combination of field criteria.