Clouds are a way of grouping a bunch of information together on a single page and defining the importance of each item based on size or a size and color combination. So, popular or important items are large and less popular or important items are relatively smaller.

Clouds are one of the defining characteristics of the whole web twenny phenomenon. Almost every site that's labeled Web 2.0 has clouds, usually in the form of tag clouds. Technorati and Flickr use clouds to show what's currently popular among their users. However, as Jeffrey Zeldman accurately opined, unstructured categorization leaves a lot to be desired.

So, when Kevin suggested that we implement a cloud view on Digg, I was a bit skeptical to start with. Weren't we just falling into the whole web twenny trend? What value would clouds serve on Digg when we've already got a nicely structured category taxonomy and you can already easily sort stories based on their popularity. Kevin and I sat down for an afternoon and worked it out and I think clouds have a place in digg that makes them uniquely useful.

What Problem Are We Trying to Solve?

We're addressing a very specific problem with clouds on Digg. In the Digg queue (the place newly submitted stories wait until they're popular enough for the homepage) the following scenario occurs frequently: a great story gets submitted and begins gaining a lot of support, lots of other stories get submitted and the first story moves down the first page quickly, the first story gets pushed off the first page, the first story is forgotten about and dies in the queue. So, in the end, a promising story never gains the popularity to get promoted to the homepage even though it may have been of high enough calibre to make it. That sucks.

So, How Do Clouds Help?

We've implemented a secondary queue view that allows diggers to see the stories in a cloud view. One of the advantages of clouds is that a lot of information can be displayed in a compact space in a meaningful way. We can display 300 stories in the queue on a single page that really isn't very tall, while still allowing quick visual scanning of the stories. Particularly with tabbed browsing, this kind of density (which means you're only seeing the titles of stories) isn't too much of a hindrance as you can open the whole story before digging it yourself. So, you can view the entire contents of the queue in a manageable space while being enabled to visually jump through only the most popular stories if you choose.

How Are We Sorting the Stories?

There are two ways we're sorting stories. As with most clouds, we're sorting by popularity. Popularity is, of course, the fundamental idea behind digging stories. So, the most popular stories are bigger and bolder and the less popular stories are smaller and more subdued. We're calculating this by relative terms — if you're browsing the "Gaming" category, for instance, the top 10% of stories are huge, the next 20% are smaller, etc. The second method of sorting is chronological. The most recent stories are at the top and the oldest stories are at the bottom.

What's So Special?

Chronological sorting might seem obvious and not very special. However, it's what adds structure to the cloud and what makes it really useful. Whereas tag clouds are often an amorphous grouping that looks cool but is hard to wrap your head around, the vertical timeline gives each item meaning even if it's not very popular. Even all of the tiny grey titles are interesting. We envision many people browsing down the entire page reading all of the stories quickly but using the sizing the consider stories more closely. Other people will quickly scan down the page and only check the biggest stories. Of particular interest will be stories of medium popularity at the top of the queue. They're not only pretty popular but they've gained that popularity quickly.

Why Don't We Simply Sort By Popularity?

You've been able to sort stories in the queue by popularity since the summer. So, you can easily view the most dugg stories in the queue that haven't made it to the homepage yet. However, this fails in two ways. First, it's only easy to browse the really popular stories that are on the brink of being promoted — say the top 30 or 40 stories. You still have to go from page to page to see them. Second, you're missing the chronological aspect of the queue. Stories are interesting not just because they have a lot of diggs, but because they received a lot of diggs very quickly. So, semi-popular items in the queue that are recently submitted are likely of more interest than more popular items in the queue that are a couple of days old.

Is this Revolutionary?

Not really. It's a slight iteration on an existing idea and may even have been implemented elsewhere I'm not aware of. It's also a somewhat intimidating way of browsing for stories that is most suitable for people who aren't complete novices to Digg. The haphazard formatting (weird line-spacing, centered text) inherent to clouds adds to this. At least for now, we're keeping it as a secondary way of viewing the queue because of these things. However, I think the combination of two important factors (time and popularity) in determining position in a cloud is an interesting idea and makes clouds really useful (and meaningful) in cases such as this.