A Better Slashdot

Anyone who regularly visits Slashdot knows that it is a unique combination of profound commentary, hilarious quips, scathing wit, and what can only be described as textual brutality. One of the first true group blogs (dating from 1997), it serves up technology related news for the geek crowd to peruse and comment on.

While it has improved over time, the formula is basically the same. Users submit articles to a group of editors. The editors revise and post appropriate articles to the site. Users can then comment on each article. A self-moderation system is used to weed the sometimes 1000+ comments under each article to a more manageable level.

Yet while Slashdot's system is very good, we've come a long way in the 8 years since its inception. So is there a way we can build a better Slashdot? With the personalization techniques seen in recommendation systems and from what we've learned of social networks, I think we can...

Goals

Let's start by listing Slashdot's strengths:

1. Articles are usually relevant (to its audience) because of filtering by editors.
2. Anyone can contribute by submitting articles or comments.
3. The plethora of commentary is diverse and interesting.
4. Each user's comments are moderated and contribute to an overall karma score which allows better contributors to float to the top.

And now its weaknesses:

1. The sheer volume of content is almost impossible to read without a full-time commitment.
   
2. Sometimes even the very best comments get buried because they are never moderated.
3. The volume of "bad" commentary outnumbers "good" commentary by at least an order of magnitude. What is "good" and "bad" is, of course, subjective.
   
4. The user moderation system is reserved to only a handful of users each day.
5. The editorial culling of articles means editorial bias.
   

I think you can boil this all down into one key problem: content overload.

To be more precise, it's not that there is too much content per se, but that there is too much content you must view in order to get to the content you want to view. Even editorial bias is just a corrallary of this problem, since if you had a way to put submissions in front of users without compromising the overall relevance of the site, you wouldn't need to use an editorial staff.

Ironically, Slashdot is probably successful for just these reasons. It just so happens to strike the current best balance between lots of content and relevant content.

That is not to say that it couldn't do even better...

Solution

Start with a user-contributed news list that allows users to discuss each article, just like Slashdot.

We first need to winnow all those articles down into something manageable. One interesting way to filter without throwing out a lot of content out of hand is personalization. We're all familiar with recommendation systems as used in online stores like Amazon and iTunes to suggest items based on the activity of other users. This approach could be applied to articles on our site in much the same manner. Article feeds could be personalized to each user based on what other, similar users like/dislike.

Assuming such a system works well, this allows us to open the article submission floodgates and depend on the personalization system to weed out spam, uninteresting or irrelevant content on a user-by-user basis. This would remove any external editorial bias yet still allow for a highly varied mix of content to coexist.

The recommendation approach could work for comments also, but we can't realistically expect at least one user rating for every comment in the system, which is a minimum requirement for such a system to work. (And besides, is a good comment necessarily a well liked one?)

Instead, we could institute a 'reputation' system similar to Slashdot's 'karma' system. Users would be able to rate comments and this would contribute to an overall reputation score of the comment's author. Unlike Slashdot, we'd open moderation to all users all the time.

Rather than hide poorly rated comments, we'd emphasize the highly rated ones using colors, larger and bolder text. This provides a positive feedback mechanism for good users, without overly punishing new or recently reformed users.

In addition, users with good reputation could receive other perks, like prefered sorting of articles that they've authored or even access to premium site features.

Of course, there is going to be a certain amount of outright spam in our articles and comments, so we ought to provide a simple spam deletion mechanism. Allowing users to suggest items as spam and having the system automatically hide them after a certain threshold is probably sufficient.

My Money Where my Mouth Is

That's a lot of words, and a lot of would's and could's. So why not step up to the plate...

Shazaaam! Webvolver.com is born.

Webvolver adheres to the design I've laid out above along with a few user-interface tweaks to make browsing comments a bit less tedious. I could go on and on, but I think the site speaks for itself, especially its FAQ page.

Obviously, as a fledging website it lacks the content to truly compete against Slashdot, but after cursory testing with friends, I feel it's certainly a great start.

So feel free to use it or abuse it and you can be the judge of whether or not it's really a better Slashdot.