Comment by Qem

2 years ago

I wonder if, society-wide, the best way to allocate resources to free/open source software would be just reducing mandatory work hours. Say, if the average ~ 8h per day were reduced to 7h or 6h, how many of those freed millions of person-hours would be redirected to hobby software projects?

8 comments

Qem

Reply
johngossman  2 years ago

You’re suggesting we reduce the work day just for software engineers (many of whom probably work longer than 8 hours) but not other engineers, teachers, autoworkers? Or for everyone, just so people have more time to work on OSS? I’m all in favor of a shorter work week, but this seems a fringe motivation

  • Qem  2 years ago

    Society-wide, for everyone. Most people would use those extra hours for things unrelated to software, like, spend more time with their children. But if even 0.1% of that goes to software, it still would be a huge amount of extra manpower. It's unfortunate despite all the productivity gains from technological advances in the last 60 years or so, the work week stayed the same size, roughly. Workers didn't earn a fair share of productivity gains.

sokoloff  2 years ago

I’d be pretty surprised if that increase averaged more than 15 seconds per person per day and shocked if the increase was more than 3 minutes on average.

  • Qem  2 years ago 

      Python 3.11.6 (main, Oct  3 2023, 00:00:00) [GCC 13.2.1 20230728 (Red Hat 13.2.1-1)] on linux
  Type "help", "copyright", "credits" or "license()" for more information.
  >>> population_base = [100_000_000, 1_000_000_000, 8_000_000_000]
  >>> for population in population_base:
  ...     print('15s/day for',population, 'people equals:', round((15*population)/3600), 'man-hours/day')
  ... 
  ...     
  15s/day for 100000000 people equals: 416667 man-hours/day
  15s/day for 1000000000 people equals: 4166667 man-hours/day
  15s/day for 8000000000 people equals: 33333333 man-hours/day

    • sokoloff  2 years ago

      Now compare that to the total productivity loss (using whatever marginal efficiency deflator you think is reasonable).

      If you're trying to get shorter workdays, go for that.

      If you're trying to get more open source labor, go for that.

      Arguing that the best way to target the second is via the first seems tenuous at best.

    • Kim_Bruning  2 years ago

      I guess that means I'm not the only person to use a repl for calculations. Neat!

      (Though this has evolved a bit for me, in order of tidiness/laziness : repl -> jupyter -> GPT+ data analysis)