The great thing about measuring developer productivity is that you can quickly identify the bad programmers. I want to tell you about the worst programmer I know, and why I fought to keep him in the team.
What evidence do one needs other than the opinion of their teammates and lead?
No one should drop players from a team due to statistics. Otherwise you’d have a non functional team of cheap wannabe-Ronaldos unable to function. Which is the reason kpi based approach fails
I literally said what evidence could be collected.
its hilarious how we spend our careers developing complex algorithms, reducing concepts into math and then delude ourselves into thinking the only algorithm that cannot be written is one that evaluates our own performance.
Because it cannot be mathematically developed. KPIs as class of algorithm are linear dimensional reductions from a complex hyperspace to a small, arbitrary reference system built on non orthogonal axes, aimed to capture non periodic, non stationary phenomena (i.e. that unpredictably evolve over time).
Mathematically, performance kpi do not make much sense for most jobs, unless the job is so straightforward that the hyperspace has such low complexity that KPIs are meaningful representation. Not even a call center job has such mathematical characteristics…
As a task, AGI is mathematically much simpler task.
However performance kpis are the only thing many have to judge, as they lack technical and personal skills to do otherwise. It’s a tradeoff, but we must recognize that kpi are oversimplifications with extreme loss of information, many time useless
I am a human, who happened to be browsing lemmy when you answered, and work in ML and with a background in algorithms and HPC. It happened to be a lucky coincidence
What evidence do one needs other than the opinion of their teammates and lead?
No one should drop players from a team due to statistics. Otherwise you’d have a non functional team of cheap wannabe-Ronaldos unable to function. Which is the reason kpi based approach fails
I literally said what evidence could be collected.
its hilarious how we spend our careers developing complex algorithms, reducing concepts into math and then delude ourselves into thinking the only algorithm that cannot be written is one that evaluates our own performance.
Because it cannot be mathematically developed. KPIs as class of algorithm are linear dimensional reductions from a complex hyperspace to a small, arbitrary reference system built on non orthogonal axes, aimed to capture non periodic, non stationary phenomena (i.e. that unpredictably evolve over time).
Mathematically, performance kpi do not make much sense for most jobs, unless the job is so straightforward that the hyperspace has such low complexity that KPIs are meaningful representation. Not even a call center job has such mathematical characteristics…
As a task, AGI is mathematically much simpler task.
However performance kpis are the only thing many have to judge, as they lack technical and personal skills to do otherwise. It’s a tradeoff, but we must recognize that kpi are oversimplifications with extreme loss of information, many time useless
That’s quite a lengthy response for the time between my post and your write up.
Not sure if you’re a bot or just used one to form your thoughts.
I am a human, who happened to be browsing lemmy when you answered, and work in ML and with a background in algorithms and HPC. It happened to be a lucky coincidence
Ok, so you’re not a software engineer. That’s good to know.
I did not say it has to be kpi on whatever other jargon you want to throw at it.
Software engineers slogan is: enough time and motivation .
Except when it comes to evaluating ourselves then it’s just not possible.
How can that be?
The bullshit excuses we tell ourselves is how we’ve ended up with ridiculous interview cycles with take home test.