An interesting project from Microsoft; reducing newsgroups, and the posters thereon, to statistics, so as to make it possible for a reader to get at the good stuff and easily ignore (or not even encounter) the dross. Quite a sound theory put forward by their pet sociologist, rendered more sound-seeming by being presented as a conclusion reached from pure statistical observation rather than as a hypothesis confirmed that way. Not sure, of course, how true that presentation is.
Unfortunate that newsgroups are a dying breed, since they're vastly superior to mailing lists for doing mailing list things; unintrusive, only getting the messages that interest you, relatively easy to pre-filter users you don't want to hear from. Perhaps this ability to filter posts automatically into a semblance of likely value-order will resurrect the beast.
The thing I really like about this technology though (and even better if it could be applied to mailing lists somehow too; certainly seems feasible for individual servers), is that if people are being rated based on their behaviour, and they know they're being rated on their behaviour - especially if ratings are accessible to all - their behaviour will probably alter, in a delicious feedback loop.
No more people just asking questions of a group and never giving anything back, no more perpetual flame wars between a few people ruining an otherwise perfectly good mailing list - the 'good' side of a flame war will not want to argue because it would ruin their reputation rating, and the 'bad' side will end up getting rated down into invisibility as they keep on arguing with whoever is new to the group. It would also provide a positive 'score' for the helpful people to feel good about themselves with.
Or! It'll result in people whining about their privacy, because people always whine about their privacy even if it's just data being constructed from other, publicly available, data.
[08:57]
|