Here's the new link:
I couldn’t resist: 3pointscircle.netlify.app

TDD as a principle was invented primarily to promote such mindset of upfront thinking and examining API surface before getting to implementation. With LLM-based assistants, conventions like TDD can fade away in favor of a text including gherkin statements of program features.

Let’s make toolings for it!

Would love to hear more about the career shift point.

Not yet. Will do when I get a chance. How about you?

Recently had a discussion about this with a friend and can’t wait for Temporal API to land.

tc39.es/proposal-tem...

Should not fix syntax errors. Imagine a scenario where formatted isn’t involved. It’s be difficult to reason why that workflow would fail as opposed to passing when formatter is involved. Not deterministic

Totally agree!

That could be a fun project

Because unless you’re experience with FSMs, it could be difficult to see the states from the get-go.

What Chris is suggesting is to start modeling with FSM, you can first list all events and their corresponding handlers, then figure out the states you need by determining what changes when those events are received.

Long walk, work out, regret life (Dec-Mar)

They can work but whether newer versions are backwards compatible with them is a choice. Understanding who uses the SDK directly impacts this decision, so that would be the basis for my reasoning.

Ah, good ol’ sparse arrays

Depends. Is there a major reason why people would want to stay on previous versions? If yes, then make migration simple and put prev versions on archive only.
From experience, when building SSKs for businesses, they tend to migrate slower and only care about migrations plan.

True! My way is to look at pictures.

Yet to find a way to follow folks I know from Twitter :)