Why Digg's new recommendation engine is a step backward

Digg's new story discovery feature is great for friends, but what about lone wolves?

Announced Monday and now live for all registered users, Digg's new recommendation engine adds a new layer of social context to the upcoming section that lists stories dugg by other users and how much their reading habits match up with yours.

Like I said earlier this week, it's a two-fold effort: one to give the social-networking element of the site some love by giving users more exposure to like-minded people, and another to make sure the site's massive influx of submitted stories gets a little more attention.

The problem is, the new system does little to solve that second problem, and in fact has taken the site a step backward from its previous version.

This no longer exists.

I speak of course of the removal of a very necessary feature called the cloud view. This would take the list of stories from just 15 a page to hundreds of headlines in a huge swarm. Better yet, those stories would be sorted in chunks (by the hour), and were set up to let you quickly eyeball stories that had begun to gain traction by headline size and color.

The cloud view has up and disappeared on us. Attempting to go to an old link with it enabled will give you an error page, and there's not a way in the user preferences to toggle it on and off. This means to go through a few hundred stories in the upcoming section, you'll need to page through at said 15 pages at a time. This wouldn't be such a big issue at a few hundred a day, but as founder Kevin Rose said Monday, we're dealing with an excess of 16,000 submissions--a number that's only getting bigger.

The fix:

The first thing that needs to be done is to bring cloud view back, but I'm almost forgetting in my nostalgia that it was a flawed system to begin with.

Cloud was great, but it was not easy on the eyes. Headlines were small, and the only way to get around that was to increase the text size in your browser.

Digg's Swarm offers an interesting prospect--tiny headlines that can expand to unveil nearly the entire entry. CNET Networks

What could make it even better is something that already exists as part of Digg labs, Digg's playground for visualizations put together by the folks at Stamen Design. One in particular, Swarmis one of the most popular and jaw-dropping cool ones on there. It tracks stories in little flying blobs. Clicking on any of these will expand it with the title, description, and current digg count. From there you can dig deeper (no pun intended) and see the actual Digg submission page with user comments and all sorts of sharing options. Thus the exploration process is complete.

Applying a similar model of swarm to cloud view is a very feasible way to make the section far more useful. People could see headlines and simply click on them to know more without having to visit another page on the site and get lost from the upcoming section entirely. Likewise, the new recommendation engine could highlight items worth looking at, without relegating them away from other submissions that might catch the user's eyeballs.

I'm afraid in the current state the upcoming section is certainly more targeted, but it's pulling users away from some of the discovery that makes Digg so fun and engaging, which is what keeps users like me coming back. Without that, it's just another popular link site.

Update: I've heard back from Digg about this, and it looks like the cloud view is not coming back. Here is the response I got:

"The cloud view was originally designed to help the Digg community parse through larger volumes of stories. As the sheer the volume of content in the Upcoming grew, now over 15,000 submissions a day, the stream became too active for cloud view to be a good user experience (which had a small number of users). We think the Recommendation Engine is the best way to filter through, and present, the most relevant content in the Upcoming section."

Makes sense, but I still think it could stick if the UI was improved. The recommendation engine will be giving some stories attention, but many smaller ones will still slip through the cracks.

Autoplay: ON Autoplay: OFF