Here's another use case perspective: consider the webcomic. There's a case where you're not searching for something—that is, every single "page" is possibly valuable content—but there's still hundreds, potentially thousands of pages to display.
The problem comes in when someone is looking for that one comic they saw a few years ago. Unlike xkcd, this comic hasn't implemented full-text search yet, so if they can't do a manual comparison sort, they have to go over every comic. What's a natural way to do this?
Still new to web development and user interface design in general; this is probably a different problem altogether.