Came up with a satisfying definition of a complex system that I should have thought of years ago.
A system is a complex system when fixing 1 bug on average causes 1 new bug to emerge somewhere else in the system.
I call it the whack-a-mole complexity threshold.
Conversation
Replying to
For those with technical interests, there are at least 2 more technical versions of this that I know of. The first is the Bode waterbed effect: improving performance in one part of the range of a control system worsens it in another part ocw.mit.edu/courses/electr
1
9
Computer science and design people seem to have a less technical waterbed theory formulation
2
8
Wrote about this stuff in a 2012 post that I probably need to revisit and
3
Replying to
I’m down as long as we maintain a clear distinction between
• complicated systems
• complex systems
1
4
Show replies
Replying to
What if every fix introduces slightly more than 1 new bug on average? Is there a way to stop the runaway trend without burning it to the ground & starting over?
1
Replying to
See my old post
1
3
Show replies
Replying to
I define it as a causality loop - cyclical nature of system means root cause analysis takes you back to your starting point. To “fix” such a system you need to encircle it, hence simultaneously fix bug at each cause in the cycle.
1
Replying to
Extremely good, I’m shocked I’ve never heard this articulated in a tweet
1
9





