It'd be interesting to see the results of the experiment - level of voting before and after, volatility of the voting before and after, rate of change of comment ordering, before and after. Are you planning on releasing some sort of aggregate data or publishing some sort of analysis?