I have first results regarding 9.06-07 sprint-task. We have 19 days of stats-history, so, I was able to get 5 days of actual karma-digits.
Here is a demo: http://s3.amazonaws.com/statsdemo/index.htm (requires flash) Formula used is: $karma = ($yesterday * 0.999) + 0.1*($avg_d_week - $avg_d_last_week) + 0.1*($avg_c_week - $avg_c_last_week) + 0.1*($avg_v_week - $avg_v_last_week); d = downloads c = comments v = rating-votes demo shows, that during these five days there is a linear declining of "karma" for all apps, but speed differs. there is a large group of apps, which have highest "karma" I have a feeling, actually, that some floating-point calculations result in precision-errors. I will try to do some tests using arbitrary precision math-library. So, my nearest plans are: * deal with precision error * make demos for other formulas If there are any proposals on formulas I will be glad to implement/test those. Probably something logarithmic? -- Alexey Zakhlestin http://www.nemein.com/ Sent from St Petersburg, Russian Federation _______________________________________________ maemo-community mailing list maemo-community@maemo.org https://lists.maemo.org/mailman/listinfo/maemo-community
