So many promises and claims in both the post and the README, yet I have no seen any evidence. I don't want to nitpick things out because it doesn't add much to the conversation, but it's assuming a lot of things about me
"the majority of developers don't give a second thought to writing code with I/O tangled in business logic"
This is a very fuzzy intro
- "the majority of developers dont do X", is something that needs to be easily verifyable
- "writing code with I/O tangled in business logic" seems like just SoC, I doubt devs have never heard about this.
The solution is a library that adds a bunch of FP with... tests? Hard bugs require session replays, not FP. FP has nothing to do with "reproducing bugs on your device". These seem like two things loosely wired together.
If you're trying to make something alike Effect-TS, then sure, this looks like a cool library, but it took a while for me to get to that conclusion.
Also, the five AI generated articles provide little value to the conversation. They don't even have a topic, other than you talking about your own library.
I have the feeling that the author is really onto something: the explicit boundary between symbolic intent and real-world execution.
Type systems don’t cut it always, you sometimes need something in-between imperative code - which is hard to test all edge cases - and pure type system chasing. The sweet spot maybe lies in having a DSL (in this case "building business mutations") and have good building blocks.
> you can assert on what the code intends to do without executing any of it
is that not just bending the meaning of mocking? Nulling things out and not executing them is a form of mocking (and the default behaviour for mocks in most languages)
So many promises and claims in both the post and the README, yet I have no seen any evidence. I don't want to nitpick things out because it doesn't add much to the conversation, but it's assuming a lot of things about me
"the majority of developers don't give a second thought to writing code with I/O tangled in business logic"
This is a very fuzzy intro
- "the majority of developers dont do X", is something that needs to be easily verifyable
- "writing code with I/O tangled in business logic" seems like just SoC, I doubt devs have never heard about this.
The solution is a library that adds a bunch of FP with... tests? Hard bugs require session replays, not FP. FP has nothing to do with "reproducing bugs on your device". These seem like two things loosely wired together.
If you're trying to make something alike Effect-TS, then sure, this looks like a cool library, but it took a while for me to get to that conclusion.
Also, the five AI generated articles provide little value to the conversation. They don't even have a topic, other than you talking about your own library.
I have the feeling that the author is really onto something: the explicit boundary between symbolic intent and real-world execution.
Type systems don’t cut it always, you sometimes need something in-between imperative code - which is hard to test all edge cases - and pure type system chasing. The sweet spot maybe lies in having a DSL (in this case "building business mutations") and have good building blocks.
Why not just use Effect? https://effect.website/
> Testing without mocking
> you can assert on what the code intends to do without executing any of it
is that not just bending the meaning of mocking? Nulling things out and not executing them is a form of mocking (and the default behaviour for mocks in most languages)