>no amount of rules or metrics can turn a bad academic into a good one...
I believe you've mistaken these systems as being for reform purposes, rather than existing as a way to weed out bad actors who want to enjoy the substantial benefits of tenure or a large salary without doing the work which justifies them.
When you reject metrics, the insidious thing is that you lack tools to catch bad actors who, due to oft-present subjectives like better than average lying ability, being liked by their team despite being largely ineffective, and exploitation of people's desire to not create conflict, the organization will never hear about them until long after the damage is done, they've made off like bandits, and you can no longer fix what they've done, since you didn't know when it happened.
Metrics cut through these specific subjectives and are therefore effectively necessary.
You can be a dirty academic, but published paper number metrics mean fakers now need to be, minimum, good enough to fake doing the job of a real academic well enough to publish, which means at least theoretically they could reform. Code reviews, pull requests, and similar programming measures do the same, leaving a clear and auditable point at which failure can be caught and later reversed. They further make it necessary to be good enough at programming to not only survive the interview, but also fake a long-term progression of source code which means to cheat you at least need to know how to program well enough to get away with it, which is within spitting distance of actually coding something useful anyways.
Admittedly, I don't think metrics create good incentives a lot of the time, but if you're thinking about the very long term and want to prevent your company's bottom from solely being a feudally structured nepotism contest, metrics are the only way to do so.
I believe you've mistaken these systems as being for reform purposes, rather than existing as a way to weed out bad actors who want to enjoy the substantial benefits of tenure or a large salary without doing the work which justifies them.
When you reject metrics, the insidious thing is that you lack tools to catch bad actors who, due to oft-present subjectives like better than average lying ability, being liked by their team despite being largely ineffective, and exploitation of people's desire to not create conflict, the organization will never hear about them until long after the damage is done, they've made off like bandits, and you can no longer fix what they've done, since you didn't know when it happened.
Metrics cut through these specific subjectives and are therefore effectively necessary.
You can be a dirty academic, but published paper number metrics mean fakers now need to be, minimum, good enough to fake doing the job of a real academic well enough to publish, which means at least theoretically they could reform. Code reviews, pull requests, and similar programming measures do the same, leaving a clear and auditable point at which failure can be caught and later reversed. They further make it necessary to be good enough at programming to not only survive the interview, but also fake a long-term progression of source code which means to cheat you at least need to know how to program well enough to get away with it, which is within spitting distance of actually coding something useful anyways.
Admittedly, I don't think metrics create good incentives a lot of the time, but if you're thinking about the very long term and want to prevent your company's bottom from solely being a feudally structured nepotism contest, metrics are the only way to do so.