problems only have one answer and often one strategy to get to the answer

Totally disagree

You’re thinking of equations, which only have one answer. There are often many possible ways to solve and tackle problems.

If you’ll permit an analogy, even though there’s “only one way” to use a hammer and nail, the overall problem of joining wood can be solved in a variety of ways.

I think this is common in scientists and researchers. They operate at the edge of knowledge with one foot in the unverifiable and their eyes peering further still into the murky unknown. There is no map nor direction where they’re going, and that extension out into the darkness is often much like superstitious belief.

What makes them different from followers of the occult that remain lost in the fog is that science returns from explorations with verifiable proof. Research extends it’s own foundations with new findings in order to venture yet another step further outwards.

IMO mathematical/logical/abstract thinking is critical for programming well, but IMO that’s different from “math degree” math.

Software as a means to an end can be used in almost every domain, so proficiency within that applicable domain is often either useful or necessary. That is to say, “math degree” math is likely needed for 3d rendering (certain games), scientific computation (incl machine learning), etc, but maybe not, otherwise. It depends on what software you’re trying to build.

To be more specific, general programming is definitely and specifically different from trig and calc. However, because math is also broad, “mathy” concepts like type theory, relational algebra, set theory are considered important for programming, even if only informally or indirectly so.

How does playing the game bring revenue? Ads?

Also, I would think that the business would be in a tougher situation if game popularity increased while tech workers weren’t around to maintain it

It’s the difference between knowing you’ll grow and graduate together with your classmates vs knowing you’re only going to see them for that one month before you move away.

I don’t understand, if you’ve got easy to delete copy-pasted code, then delete it. It’ll be a nice and cathartic exercise.

But sounds like what you’re really talking about is code that isn’t easy to delete.

Distributing power across a group of communities over the same topic (e.g. like seats in a congress/parliament) is a nice thought.

However, my second thought was how vulnerable that is in a fediverse. To continue the analogy, an adversary could create new states (server/communities) of arbitrary population (accounts) at will.

Whiskey

IMO folding to hide is about equivalent to moving all contents to another file/private function:

def bad_function(args):
    return _hide_elsewhere(args)

i.e. does nothing. Real solution to pyramids of doom is to fix the code.

That’s changing the goal posts to “not static”

Sounds easy to simplify:

Use one of: constructor A(d), function a(d), or method d.a() to construct A’s.

B and C never change, so I invoke YAGNI and hardcode them in this one and only place, abstracting them away entirely.

No factories, no dependency injection frameworks.

IMO factory functions are totally fine – I hesitate to even give them a special name b/c functions that can return an object are not special.

However I think good use cases for Factory classes (and long-lived stateful instances of) are scarce, often being better served using other constructs.

Dedicated incremental static type checkers for dynamic languages already exist. In particular, Pyright for Python is fantastic and in many ways surpasses the type systems of classic typed languages

But would you pay for it?

My employer’s paying for my access, and I only find it a bit useful here and there

Maybe my company gets a great discount or something, but if they would pay me the subscription cost to give up Copilot, I wouldn’t miss it

You can reference envs from the host in docker compose, so code it in instead of manually passing tribal knowledge in: https://stackoverflow.com/a/73826410

Simpler to keep everything in one compose file if you can, under a test service that doesn’t build unless explicitly named

Un-weird that env var and use the normal, boring feature of defining environment under your test service

I’ve often been able to alias drun='docker compose run --rm --build' and simplify down to:

drun test

Should be able to encode all those wayward args into docker-compose.yml or Dockerfile and only use vanilla docker commands – that’s the whole point of containerization

In the US? IMO only possible in exclusive environments similar to saunas at spas or membership-based clubs/gyms

I think your ideas are too non-practical/specialized/advanced/low-level for your stated goal of 'digital literacy". They read more like college intro/followup course material and are too esoteric to be readily absorbed, esp by generic teenagers, even if they’ve self-selected to be “lightly interested”.

One of the best tutorials on really “grokking” git concepts, and it’s online and interactive: https://learngitbranching.js.org

For programming, start with buildings things for yourself. Be practical, start small, and iterate, regardless if you consider the previous iteration was a success or failure. I’ve heard good things about https://automatetheboringstuff.com/ (in Python) in this regard.