Infinite scroll is like finishing a sandwich, and the server plops another one in front of you without asking what you want on it, or if you want it at all. If you’re full, or you don’t like what they chose? Too bad, it’s on your plate now! To make matters worse, sometimes if you put the sandwich down for a moment to eat some chips, they’ll think you’re done and swap your sandwich for a different one!

Slate just replaced pagination with infinite scroll on their articles. Yes, pagination sucks. A multi-page article on the web is like a burger that’s been sliced up into wedges, and you only get one wedge at a time, forcing you to go back to the counter every few bites. But infinite scroll isn’t an improvement.

Both approaches impose the wrong structure on a single unit. Search results and timelines are one thing, but for an individual piece of content, the best way to map it to a web page…is to just map it to a web page.

Update (Sep 2016): Combined with giant images and complex layouts that slow down browser rendering (*cough* CBR), it’s even worse. To continue the lunch analogy:

1. You order a sandwich with a cup of soup and a side salad.
2. After an interminable wait, you get the sandwich and soup, but no salad, and no spoon. The waiter rushes off before you can say anything.
3. Eventually you’re able to flag someone down and ask for the spoon and the salad.
4. You munch on the sandwich by itself, which is at least a decent sandwich.
5. Finally the waiter comes back with a whole pizza, and takes away your half-eaten sandwich.
6. You still don’t have a spoon, but that doesn’t matter because the waiter took the soup too.