Good Works, Bad Works, but Karma doesn’t work
Richard McManus at ZDNet’s Web 2.0 Explorer and Read/Write Web investigates the karma system at Reddit, and wonders if it’s better than Digg’s:
[from » Collaborative filtering: comparing Reddit’s karma system to Digg]
Indeed Reddit’s Alexis Ohanian says that “with reddit, we’re hoping that by focusing on filtering, users will be inclined to vote up links that genuinely interest them”. The Reddit method then is trying to capture that elusive social software principle of getting the user to reward him or herself first and foremost, but actually the system is enhanced at the same time. As Alexis said, “The nice thing about this is that although users are serving themselves by voting to train a personal filter, the by-product of their honesty is that the community gets a more accurate idea of what’s really popular.”
What do you think. Is Reddit’s karma system a better - more honest - way to rank stories and users than Digg’s? Or do you think Digg has the right approach, but just needs to address the groupthink and spam issues that come with scaling to thousands of users?
My strong belief is that these rating/ranking filters are failing at a fundamental level: not because they don’t work, or because they can be gamed, but because they are
- closed, and
- tightly integrated to a specific solution.
Each application could influence this foundational karma in dissimilar ways. On one application, say Oyogi, your karma would increase by correcting answering questions directed to you, while Technorati might increase your karma based on the number of inbound links to your blog, and Typepad might increase your karma based on the number of positive comments you receive on your blog.
Since we are in the Web 2.0 era of mash-ups, won’t someone break their karma system out, and set it up in this way, so that we can avoid a fragmented collection of non-interoperable solutions? People who would like to have various, separate identities could still do so, and the solution could support anonymity, but the benefits of a open karma solution, that all could draw upon, is so obvious that I am actually surprised it hasn’t been done already.

