Piglei
Using import-linter to
- define architectural layers
- check to make sure imports don’t violate (import from upper layers)
- can also check for more contracts, such as
  - forbidden - disallow a specific from/to import
  - independence - list of modules that shouldn’t import from each other
Fixing violations
- a process introduced to set exceptions for each violation in a config file
- then fix violations 1 at a time (nice approach)
- use the whole team if you can
Common methods for fixing dependency issues
- Merging and splitting modules
- Dependency Injection, including using protocols to keep type hints without the need to import just for types
- Use simpler dependency types
- Delaying function implementations
  - module global methods set by caller, or adding a simple plugin/callback system
- Configuration driven
  - Setting import statements in a config file and using import_string() at runtime
- Replace function calls with event-driven approaches

Michael #2: Mountaineer

Brian #3: Why Python's Integer Division Floors

Guido van Rossum
Integer division always floors (toward negative infinity) instead of truncating. (C truncates)
5//2 → 2
-5//2 → -3
5//-2 → -3
Reason,
- For nice mathematical relationships with // and % (modulo).
- a//b = quotient (q), a%b = remainder (r)
- such that b*q + r = a, and 0 <= r < b
  - This works for both positive and negative a values
  - For negative b, the second rule has to change to 0 >= r > b
If you truncate (like C does), you have to use abs(r) for the first rule to work.
Theory of why C doesn’t do it this way: Probably a hardware limitation at the time when C was designed, due to “sign + magnitude” integers instead of modern two’s compliment integers.

Michael #4: Hatchet

Hatchet is a distributed, fault-tolerant task queue which replaces traditional message brokers and pub/sub systems.
It’s built to solve problems like concurrency, fairness, and durability
Concurrency, Fairness, and Rate limiting: Enable FIFO, LIFO, Round Robin, and Priority Queues with built-in strategies to avoid common pitfalls.
Architected for Resiliency: Customizable retry policies and built-in error handling to recover from transient failures.

Extras

Brian: