That's both perfect for this thread and terrible in general. If you can see the analogy between text and pictures, it drives the point exactly the right way: i

That's both perfect for this thread and terrible in general. If you can see the analogy between text and pictures, it drives the point exactly the right way: in both cases you expect a database to know things it either can't or forgot. If it had a good picture of the zoomed in background it could probably generate a very good representation of what the cropped part would look like; same thing works with text.