Richard's proposition is basically that metrics should be created and used because its always better to have more data. This is only true if the data is good. T

Richard's proposition is basically that metrics should be created and used because its always better to have more data. This is only true if the data is good. The biggest problem I can see is that metrics are easy to fake. That's fine on sites StackOverflow where the stakes are very low but when metrics are accorded extremely high value, as is being proposed here, you immediately create a huge incentive to pollute the data. The only way out of this is to create some kind of quality assurance of the metrics themselves. To me that seems like a monster problem that no startup can possibly handle. If they restrict their scope a bit more than trying to liberate all of academia from the publishers there are probably ways to add value.