Thursday, April 08, 2021
Subtracting from knowledge about subtraction
Another game-theory "study" on a well-known and important topic.
When trying to solve a problem or fix a machine, people tend to add more bells and whistles instead of removing pieces. It's a definite tendency, and it takes serious discipline to reverse the tendency.
I have to slap myself upside the head when I'm debugging a program. SIMPLIFY, SIMPLIFY, SIMPLIFY!!! It's always "easy" to add a conditional or a try/except pair that solves the APPARENT problem in a hard-wired way. This solution is nearly always wrong, and nearly always creates more bugs.
[Example: You see x=0 sometimes in a situation where it should be physically impossible for x to reach 0. The add-on "solution" is something like this: If x=0, set x=0.001. Instead, you should dig in and figure out why the physically impossible situation is showing up.]
The article quotes one observer who understands the tendency clearly:
Parkinson. Bureaucracies ALWAYS grow. Spending ALWAYS increases, even after the "problem" has entirely disappeared.
BUT:
The article misses one important category of bureaucratic problems where subtraction is PREFERRED. When we're thinking about cleaning a room or an environment, we sternly resist adding something. Chemical pollution of air and water is often EASILY fixed by adding a plant or a bacterium or a reactant or an absorbent powder, but agencies refuse to adopt solutions in that direction. They insist that the usable solution must be purely subtractive.
So this study subtracts from existing knowledge about subtraction.
Another game-theory "study" on a well-known and important topic.
When trying to solve a problem or fix a machine, people tend to add more bells and whistles instead of removing pieces. It's a definite tendency, and it takes serious discipline to reverse the tendency.
I have to slap myself upside the head when I'm debugging a program. SIMPLIFY, SIMPLIFY, SIMPLIFY!!! It's always "easy" to add a conditional or a try/except pair that solves the APPARENT problem in a hard-wired way. This solution is nearly always wrong, and nearly always creates more bugs.
[Example: You see x=0 sometimes in a situation where it should be physically impossible for x to reach 0. The add-on "solution" is something like this: If x=0, set x=0.001. Instead, you should dig in and figure out why the physically impossible situation is showing up.]
The article quotes one observer who understands the tendency clearly:
But curbing our love of excess will take more than nudges and a clear mind, says Hal Arkes, a judgement and decision-making researcher at Ohio State University who was not involved with the study. Organizational and political leaders, especially, abhor cutting the fat. “If you add more people and more dollars, you won’t make any enemies, you’ll just make friends,” Arkes says. “Subtraction has serious downsides.”
Parkinson. Bureaucracies ALWAYS grow. Spending ALWAYS increases, even after the "problem" has entirely disappeared.
BUT:
The article misses one important category of bureaucratic problems where subtraction is PREFERRED. When we're thinking about cleaning a room or an environment, we sternly resist adding something. Chemical pollution of air and water is often EASILY fixed by adding a plant or a bacterium or a reactant or an absorbent powder, but agencies refuse to adopt solutions in that direction. They insist that the usable solution must be purely subtractive.
So this study subtracts from existing knowledge about subtraction.
Labels: Parkinson
¶ 8:51 AM
