The purge is petering out. I zapped just a couple of non-trolls, though I did focus quite a bit on ratings abusers. Remember when I said this would be arbitrary at times? I pulled up a random survey of comments that had both recommends and hides. I looked to see if there was a genuine reason for both sides to do what they did. If not, I pulled ratings from the side that was in the wrong.
Most of the time, it was people trying to hide non-trollish or objectionable comments. A hide rate is not "I disagree with your premise". But of course, the true purpose of the hide rate is often ignored. In any case, I didn't pull ratings for everyone who has abused ratings, just for a random sample. Please try to use ratings properly!
But all of this ratings stuff is not long for this world anyway. The team and I spent some time this morning doing some more sketching of our future CSMAS (Community Self-Moderation Adjudication System), and we came to a realization -- we don't need mojo.
The new system will refer problem comments to a site-wide adjudication process that will culminate with a vote. Yes, we are considering the possibility that cliques will try to game this, and are accounting for that as we sketch that out.
If someone thinks a comment actually should be hidden (instead of simply eliciting disagreement), she can hit "flag" (or whatever we end up calling it). Unlike the current system which requires a load of hides to impact a user's mojo, and even more hides if the comment is uprated, this system would merely require a few flags to kick it into the adjudication system. The flagger will also be required to state why she thinks the comment should be removed -- a sort of opening statement in what will essentially be a trial.
Now here's the thing -- if a comment gets flagged for adjudication, someone will get punished. If the community decides the comment was hide-worthy, then the commenter will lose privileges for X amount of time, where X increases exponentially for each violation. If the community decides the comment wasn't hide worthy, then those flagging the comment will lose privileges. Each case, kicked into adjudication, will allow for debate via comment thread, so people can render their opinions on the matter before casting their vote.
And yes, there will be a court of appeals after the community renders it's decision -- me.
Right now, we use mojo and TU for two reasons -- to give people the ability to hide rate, and to create new groups. But since hide rating is going away, it removes that reason for having TU. And there are better ways to filter for group creation. TU can then be replaced with a system of badges to give people credit for the specific actions they excel at, whether it's commenting, or being good community members, or posting diaries/stories/essays/etc, or whatever.
This is a fairly radical reworking of community reputation. I just don't think the number of bars is really reflective of the value of someone to the community. And saying someone is "trusted" while someone is not is kind of shitty.
I often remind people that I love to experiment, and I'm not afraid to try new things. As we sketch out this new CSMAS, we're finding that it's impacting the very core of how we measure community reputation, and I'm finding myself more ready than ever to reconsider how it works.
We've had a couple of weeks of fantastic meta discussion (and I do love it!), and I want to keep that going. Your feedback plays a seriously big role in how this is shaping up. So keep it up!
On another note, the development team is dedicating next week to minor site fixes. Among them, a couple should prove popular -- you'll be able to collapse the diary boxes on the right-hand corner on the front page, and have your browser remember your decision during future visits. So if you only care about recent diaries, you'll be able to collapse both the Community Spotlight and Recommended lists, and have far less scrolling to get to what you really want. In the future, we'll even allow re-ordering of content on the front page, but that won't happen until we transition full away from Perl to Rails, which should be a year-long process. (Rails and Perl are underlying frameworks upon which the site code is built. Perl is a bit geriatric, while Rails is the hip and modern.)
We're also going to move the Welcome Back box to below the medium rectangle ad on the right column, fix the auto-tip-jar placement bug, add mouse-over intro text to all diary headlines, and allow users to follow themselves.
Then we're going to go back to working on a couple of bigger projects, like the CSMAS (hurray for horrible acronyms!) and image upload projects that are going to be at least a couple of months away from completion.