3 ms·

> 0.1 Bellard Off topic, but a log scale might be useful: 0.1 Bellard --> -10 deciBellards. That allows for: 0.001 Bellard --> -30 deciBellards. Problem: Prog

> 0.1 Bellard Off topic, but a log scale might be useful: 0.1 Bellard --> -10 deciBellards. That allows for: 0.001 Bellard --> -30 deciBellards. Problem: Programmers with negative productivity cannot be represented on the same log scale.

araes 3y ago
Sure they can. Abs(). Or if you prefer to not have quite so much micro-nano, you could also use Cumulative Distribution Functions [1] which are basically just sums of Probability Density [2]. Are they a 4σ programmer, 1σ programmer, -0.5σ programmer, -2σ programmer? Plus, most people are "average" not negative productivity, and CDFs let you use really fun stuff like Beta Distributions (variable, shapable distributions) and Gamma Distributions (exponential distributions). They're super sweet as far as probability statistics. [1] https://en.wikipedia.org/wiki/Cumulative_distribution_function https://en.wikipedia.org/wiki/Cumulative_distribution_functi... [2] https://en.wikipedia.org/wiki/Probability_density_function https://en.wikipedia.org/wiki/Probability_density_function [3] https://en.wikipedia.org/wiki/Beta_distribution https://en.wikipedia.org/wiki/Beta_distribution [4] https://en.wikipedia.org/wiki/Gamma_distribution https://en.wikipedia.org/wiki/Gamma_distribution
- hedora 3y ago
  The number of people that are average is infinitesimal, even if you define average as “mode” instead of median or mean. About half are better than median though.
analog31 3y ago
>>> Problem: Programmers with negative productivity cannot be represented on the same log scale. This is similar to the problem of price-to-earnings ratio. The ratio goes asymptotic as earnings goes through zero. It would be better to quote earnings-to-price ratio. Another screwy reciprocal unit is miles per gallon for cars.
- hedora 3y ago
  Now that I have an EV with 130 miles of range, I’m finding miles per kwh to be much more useful than efficiency. I’d bet range anxiety was also a thing for early gasoline powered cars, so the early adopters of those probably preferred mpg over gpm.

araes 3y ago

Sure they can. Abs(). Or if you prefer to not have quite so much micro-nano, you could also use Cumulative Distribution Functions [1] which are basically just sums of Probability Density [2]. Are they a 4σ programmer, 1σ programmer, -0.5σ programmer, -2σ programmer? Plus, most people are "average" not negative productivity, and CDFs let you use really fun stuff like Beta Distributions (variable, shapable distributions) and Gamma Distributions (exponential distributions). They're super sweet as far as probability statistics. [1] https://en.wikipedia.org/wiki/Cumulative_distribution_function https://en.wikipedia.org/wiki/Cumulative_distribution_functi... [2] https://en.wikipedia.org/wiki/Probability_density_function https://en.wikipedia.org/wiki/Probability_density_function [3] https://en.wikipedia.org/wiki/Beta_distribution https://en.wikipedia.org/wiki/Beta_distribution [4] https://en.wikipedia.org/wiki/Gamma_distribution https://en.wikipedia.org/wiki/Gamma_distribution

hedora 3y ago
The number of people that are average is infinitesimal, even if you define average as “mode” instead of median or mean. About half are better than median though.

analog31 3y ago

>>> Problem: Programmers with negative productivity cannot be represented on the same log scale. This is similar to the problem of price-to-earnings ratio. The ratio goes asymptotic as earnings goes through zero. It would be better to quote earnings-to-price ratio. Another screwy reciprocal unit is miles per gallon for cars.

hedora 3y ago
Now that I have an EV with 130 miles of range, I’m finding miles per kwh to be much more useful than efficiency. I’d bet range anxiety was also a thing for early gasoline powered cars, so the early adopters of those probably preferred mpg over gpm.