Chunk Size
It's a Trade Off
Me: cut your food into chunks that are easy to chew
Them: but if I keep cutting forever I end up splitting atoms and then there’s nothing left to eat
Me:
My “mind-sized chunks” tweet got a bunch of replies of the form, “Yeah, but if you make the chunks too small then the system is harder to understand, not easier.” My question: what is so threatening about “mind-sized chunks” that “smart” people would immediately jump to reductio ad absurdum.
Having come up in Smalltalk, I’m willing to believe that my “small” is smaller than many folks today. Still, there’s a big space of possibilities out there and the one point you’re sitting on now isn’t sacred.
It’s a trade off, folks. Smaller chunks cost in exchange for benefits. Bigger chunks cost in exchange for benefits.
"What do you mean I don't have the mental capacity to understand these chunks and need to make them smaller?!?!"
So, when you posted this on Twitter I replied with "Also Sprach Raymond Hattinger" and a link to a video.
Raymond is a core Python contribuitor and in this talk he argues that humans have a limited set of registers.
Just like machines, the possibility of holding more data that the registers, can only be done if we either chunk data together and/or use aliases.
Similarly, in the book, "Team Topologies", the authors talk about cognitive load for teams as a way to define the border at which we should a team can expand its domain. It takes this idea from Evans on DDD.
Here's the video for the first example.
https://www.youtube.com/watch?v=UANN2Eu6ZnM