- A log scale is far better than a linear one. While it's not top-end limited, it does tend to flatten the scale markedly.
- A decay function on older contributions is helpful. Especially if users start abusing the system. Bank karma, abuse system, run karma down slowly, with an eternal linear system. With a logrithmic decaying system, abuse degrades reputation far more rapidly.
- Other options are to rank users by percentile or other bases. Given the various problems of innate / explicit moderation / ranking (see A Better Wayto Rate Films: http://blog.goodfil.ms/blog/2011/10/07/a-better-way-to-rate-... ) an explicit percentile norming preserves meaningful distinctions (though individuals ratings may not be directly comparable).
I don't think encouraging activity is necessarily good. A lot of comments do not really contribute something useful to the discussion. I'd rather people just posted when they have something really useful or insightful to say. Less comments means less time wasted for the readers.
Actually this isn't very accurate. There are several posts higher than that, but they don't show up. Here are a few I remember being higher, I'm sure there are more. Granted one of them was about Steve Jobs too.
While this is obviously the post with the most points, I just noticed that something might be broken with this. Looking for posts with over 1000 points omits the following post with 1770 points:
There are several other high scoring posts that don't show up by that method, because it's meant as a filtering tool for stories on the frontpage and one page deeper. Obviously these stories tend to be far apart so don't see them all with that /over method, only recent ones that haven't fallen off the first two pages yet.
