Nat (she/they)

libraries let them in

The feeling of not being able to breath properly when I inhale there is pretty unpleasant imo.

It’s a neat option, but should not be forced.

Websites should not get to dictate my security model. I’ll accept annoying me about being less secure because I get that people are dumb, but you’ve gotta choose somehow! Also, any passkey is safer than a password, so that’s still BS.

It probably comes from C, which also copies by default, though it doesn’t have copy constructors.

Not if Random writes to global state, that’s a side effect that must be preserved

Wait till you find out how the runtime manages multiple concurrent timers

Okay, so it’s just a distributed linked list where earlier entries can’t be changed without changing everything up to the head? I guess I can see a few niche uses for that. In my head I was just thinking “surely that can’t be it” because it’s so simple, hence thinking the proof-of-[x] thing would also have to be part of it.

I don’t see how a blockchain is necessary there. Couldn’t they just use transaction databases and simple messaging between banks? Also, what is your definition of blockchain? Just a distributed linked list? Proof of work (the part I don’t understand the need for)?

you don’t, and that’s good, but that doesn’t make it universal. as for the split between these 2 types, idk so I won’t even guess

I’ve been using primarily webp for like half a decade and I haven’t noticed many compatibility issues or bad quality. I guess if your software hasn’t been updated in the past decade it won’t work, but in that case I guess we should never make a new image format again?

/dev/null is web scale, it maintains sub 1ms times no matter how much load you give it!

That’s great until someone says “I’m not being a dick, I’m just telling the truth” while being a dick. This is especially easy to pull off against minorities, because the aforementioned “truth” can be based on stereotypes or inaccurate media portrayals.

Dead code elimination but with a different name for some reason

I used them heavily, but I still run into issues of guessing what type library functions return or expect me to pass in. Sometimes there was no answer because the authors I guess wanted to be cool and accept any type that kinda fits, or they return either this or that type based on the arguments and now I have to assert which one it is to the type linter. And then I’d still get runtime errors about failed property accesses deep in library code and have to figure out wtf they wanted me to do.

Eh, Python has a very small slice where I’d consider it the right tool for the job for me. It’s for when I want a less awful bash script, but going much bigger than that makes me miss type systems.

okay, let me elaborate: the hardware acceleration requires custom software to interact with it yes, BUT, they probably still use ffmpeg and just pass -vcodec custom_thing to it

ffmpeg codec

C’s compiler prevents common type bugs and handles things like register allocation for you? So does skill.

A null pointer exception is technically memory safe, you can get equivalent behavior with .unwrap() on an Option in Rust.