DEV Community: Jordan Bravo

When Not To Use Python

Jordan Bravo — Mon, 08 Jan 2024 15:18:14 +0000

Python is useful in some specific situations:

Scripting
Wrangling data
Large number of ML libraries with Python bindings to underlying C/C++

It is a terrible choice for backend web services in enterprise environments across a team.

Hypothetically it's a good fit for startups that need to quickly hack together a minimum viable product. The problem is that all too often there is no time to rewrite it as the company grows. Python's design does not discourage bad habits and in fact promotes some of the bad habits.

Problems with Python:

Significant whitespace
Slow
Dynamic typing
Async is an afterthought
Python 2 to Python 3, broken APIs, essentially a new language
Installation and environment is a headache:
- Installs globally to system
Global variables
Mutable by default
Implicit export (everything is exported)
Imports/namespacing
Heavily leans into object oriented code
Allows shadowing and overwriting, and variable declarations don't use a keyword. This might seem like a good feature at first because it's less typing, but it's actually terrible when working in a large complex codebase. Example: I write 'data = xyz', but elsewhere in the code there exists 'data = abc', I will have no idea that I just overrode a variable.

ASDF is a single version manager for all languages

Jordan Bravo — Sat, 01 Apr 2023 15:22:41 +0000

I had to manage multiple versions of NodeJS, so I installed Node Version Manager.

Then I had to also manage multiple versions of Python, so I installed PyEnv.

Then I had to manage multiple versions of Rust, Golang, Lua...

There must be a better way.

Enter asdf, a single version manager for all programming languages.

Homepage: https://asdf-vm.com/

Caveat: It’s currently only available for MacOS and Linux.

How to speed up Docker on MacOS

Jordan Bravo — Mon, 27 Mar 2023 15:33:53 +0000

Problem: Containerization is a feature of the Linux kernel (namespaces + cgroups + chroot) and so MacOS and Windows must run a virtualized Linux environment to use Docker, which has a significant performance penalty.

Solution: To help mitigate this issue, Docker (the company) has some recommended configuration tweaks for optimizing performance on MacOS:

Step 1. In Docker Desktop Settings > General > Choose file sharing implementation for your containers. Select VirtioFS and restart Docker Desktop.

Step 2. If you’re using any bind mounts, add the delegated option when you specify the mount source and destination in the docker-compose.yml file, for example:

  my-service:
    image: my-image
    container_name: my-container
    volumes:
      - ./my-app:/app:delegated  # this is the relevant setting

References:
https://www.docker.com/blog/speed-boost-achievement-unlocked-on-docker-desktop-4-6-for-mac/
https://docs.docker.com.xy2401.com/compose/compose-file/#caching-options-for-volume-mounts-docker-for-mac

DEV Community: Jordan Bravo

When Not To Use Python

- Installs globally to system

ASDF is a single version manager for all languages

How to speed up Docker on MacOS