The Wymanator

Here's a classic case of a user interface designed by someone who thinks like a programmer as opposed to an end user:
Now to a programmer this is an intuitively obvious interface. Google's back end supports the ability to display search results starting with any random page, so why not expose this capability to the end user?

Unfortunately, from the end user's perspective, this makes no sense whatsoever. Google presents its search results ordered by how likely they are to be of interest to the user. If a user doesn't find what they need on page 1 they typically will follow one of two courses of action:

• Re-phrase their search terms and repeat the search (the approach I use 90% of the time)
• Go to page 2 to view the next most relevant set of search results

No user in their right mind would say to themselves "Hmmm, I didn't find what I was looking for on page 1. Let me see what's on page 4." But this is exactly what the Google interface supports.

The only navigation controls needed here are Next, Previous and First page.

Microsoft's Windows Live search (boy do I hate that name), does an excellent job of finessing this whole issue by supporting an "infinite scroll" feature. As the user scrolls through the result set using either the mouse wheel or scroll bar, the next search result set is automatically retrieved under the covers. This is very much analogous to how Google Maps supports panning which, compared to other map services which required page reloads to effect panning, represented a quantum leap in the interface.

Of course, Google's search thrives on light weight and speed -- something which has never concerned Microsoft -- so it's not clear if it makes sense for Google to apply heavy weight Ajax technology to their bread and butter.