#317: Most loved and most dreaded dev tools of 2022

Published Tue, Jan 3, 2023, recorded Tue, Jan 3, 2023

Watch the live stream replay

About the show

Sponsored by Microsoft for Startups Founders Hub.

Connect with the hosts

Michael #1: StackOverflow 2022 Developer Survey

Last year we saw Git as a fundamental tool to being a developer. This year it appears that Docker is becoming a similar fundamental tool for Professional Developers, increasing from 55% to 69%.
Language: Rust is […] the most loved language with 87% of developers saying they want to continue using it.
JS Frameworks: Angular.js is in its third year as the most dreaded.
Let me Google that for you: 62% of all respondents spend more than 30 minutes a day searching for answers or solutions to problems. 25% spending more than an hour each day.
The demise of the full-stack developer is overrated.
I do wish there were more women in the field.
Databases: Postgres is #1 and MongoDB is still going strong.
The “which web framework do you use?” question is a full on train wreck. Why is this so hard for people to write the question? Node.js or Express (built on Node) vs. FastAPI or Flask (but no Python?)
Most wanted / loved language is Rust (wanted) and Python/Rust tied for most wanted.
Worked with vs. want to work with has some interesting graphics.

Brian #2: PePy.tech - PyPI download stats with package version breakdown

Petru Rares Sincraian
We’ve discussed pypistats.org before, which highlights
- daily downloads
- downloads per major/minor Python version
- downloads per OS
PyPy is a bit more useful for me
- default shows last few versions and total for this major version
- “select versions” box is editable.
  - clicking in it shows dropdown with downloads per version already there
  - you can add * for graph of total
  - or other major versions if you want to compare
- daily/weekly/monthly is nice, to round out some noise and see larger trends
- Oddity I noticed - daily graph isn’t the same dates as the table.
  - off by a day on both sides
  - not a big deal, but I notice these kinds of things.

Michael #3: Codon Python Compiler

via Jeff Hutchins and Abdulaziz Alqasem
A high-performance, zero-overhead, extensible Python compiler using LLVM
You can scale performance and produce executables, even when using third party libraries such as matplotlib.
It also supports writing and executing GPU kernels, which is an interesting feature.
See how it works at exaloop.io
BTW, really terrible licensing.
- Free for non-commercial (great)
- “Contact us” for commercial use (it’s fine to charge, but give us a price)

Brian #4: 8 Levels of Using Type Hints in Python

Type Hints for Basic Data Types.
- x: int
Define a Constant Using Final Type
- DB: Final = '``PostgreSQL'
- (ok. I haven’t used this one at all yet)
Adding multipe type hints to one variable.
- int | None
Using general type hints.
- def func(nums: Iterable)
- Also using Optional
Type hints for functions
- def func(name: str) → str:
- (I probably would put this at #2)
Alias of type hints (not used this yet, but looks cool) PostsType = dict[int, str]

new_posts: PostsType = {1: 'Python Type Hints', 2: 'Python Tricks'}
Type hints for a class itself, i.e. Self type from typing import Self

class ListNode: def __init__(self, prev_node: Self) -> None: pass
Provide literals for a variable. (not used this yet, but looks cool) from typing import Literal weekend_day: Literal['Saturday', 'Sunday'] weekend_day = 'Saturday' weekend_day = 'Monday' # will by a type error

Extras

Brian:

Michael:

Joke: vim switch

Full transcript