We need database joins so that we can derive new, interesting propositions from the data already in the database. A database is a set of facts about something.

We need database joins so that we can derive new, interesting propositions from the data already in the database. A database is a set of facts about something. Joins are the logic reasoning that lets you deduce new things. If you never need to find out new things, then you don't need joins. So you can get away without joins when you have a simple system that is doing repetitive work. You can't avoid them when you are doing complex analysis or presenting data in a variety of different ways.