Doing It Wrong

learn in public

You are here: Home / 2024 / Archives for July 2024

Archives for July 2024

The Programmer’s Brain

posted on

Notes on The Programmer’s Brain, by Felienne Hermans

  • Storage Strength (SS) & Retrieval Strength (RS)
    • SS: how well something is store in Long-Term Memory (LTM)
    • RS: how easy it is to remember something
  • it is the act of trying to remember (w/o additional study) that improves RS
    • this is why flashcards work so well
  • LTM is stored (and to a degree, filtered through) schemata
    • I bet this is what makes a “memory palace” / the Method of Loci so effective
  • elaborate new information (actively think about, relate to existing knowledge) to strengthen RS
  • Working Memory (WM) vs Short-Term Memory (STM) vs LTM
    • WM: like a processor—STM applied to a problem
    • STM: like Random Access Memory (RAM)
    • LTM: like Solid State Drive (SSD) or Hard Disk Drive (HDD)
  • moving from shallow to deeper understanding of a codebase
    • find a focal point (e.g., main.js)
    • expand knowledge from focal point
    • understand a concept from a set of related entities
    • understand concepts across multiple entities
  • frameworks that, for example, auto-wire up dependency injection, can cause a lock of Plan Knowledge even though you understand the code (Text Knowledge)
  • language abilities, WM capacity, and reasoning skills have a greater impact on a programmer’s end result than numeracy (the ability to understand, reason with, and apply simple numerical concepts)
    • this is also true for your learning rate

Thoughts on Microservices

posted on

Thoughts on microservices the first time I had to structure code so that it worked well on a physical boundary and poorly on its logical boundaries.

Because of cold-start times and serverless ideology, we keep things small in microservices. We miss out on some stuff, I think. Microservices are the hard edge of the API interface, everywhere. Services (maybe just domain-oriented ones) allow you the tried-and-true wisdom of grouping together those things that share a reason to change. You can find common patterns, optimize things, share like code yet develop a good sense of what shouldn’t be abstracted. Then, you expose what you want to the world with the hard edge of the API interface, where you get what you get–what the API design allows.

I was just writing all this when I found Towards Modern Development of Cloud Applications. The paper seems fantastic–though I’ll admit that I hope to see it all done in WASM or something. I should also admit that I probably don’t understand it all that well. I’d love to mess around with it until I do.