412 karma · joined May 9, 2023
Prepare to have your mind blown: https://ratfactor.com/forth/the_programming_language_that_wr...
I think we forget the big picture that "no one gets out alive" and that it's up to each of us to spend the time we have in a way that makes us happy. If you can afford to not spend 1/3 of your time at a job in order to fund the remaining 2/3 of your existence, lucky you! But if not, choosing to slack off at a job you hate is essentially choosing to throw 1/3 of your life away being miserable. (Not judging if that's your choice, but its baffling to me.)
Now let's take "the future" part... you seem to be impugning Async Rust for something it's not even purported to do in the present. What's the point of this?
You found a bug in monoio it seems... I don't see the argument you've presented as supporting the premise that "Async Rust is not safe".
Yes, it prevents many useful programs.
I think it also prevents many many many more useless broken incorrect programs from wreaking havoc or being used as exploit delivery vehicles.
Are you suggesting your customers don't care about CVE's, even indirectly when it affects them?
Yet another reason I have a hard time trusting software that is the product of integrating many disparate pieces with different engineering processes into something used in any life-safety or mission-critical application.
> a revolutionary data storage format that can survive for billions of years
Clearly it hasn't been verified to be able to survive for billions of years, but it is thought to be able to survive for billions of years based on extrapolation from tests done over time scales multiple orders of magnitude smaller.
https://www.fastly.com/blog/measuring-quic-vs-tcp-computatio... has good details and was done almost five years ago.
The solution isn't in more UDP offload optimizations as there aren't any semantics in UDP that are expensive other than the quantity and frequency of datagrams to be processed in the context of the QUIC protocol that uses UDP as a transport. QUIC's state machine needs to see every UDP datagram carrying QUIC protocol messages in order to move forward. Just like was done for TCP offload more than twenty years ago, portions of QUIC state need to move and be maintained in hardware to prevent the host from having to see so many high-frequency tiny state updates messages.
What is needed is for QUIC offload to be invented/supported by HW so that most of the high-frequency/tiny-packet processing happens there, just as it does today for TCP offload. TCP large-send and large-receive offload is what is responsible for all the CPU savings as the application deals in 64KB or larger send/receives and the segmentation and receive coalescing all happen in hardware before an interrupt is even generated to involve the kernel, let alone userland.
I love this marketing soundbite too:
> “reMarkable gives me the deep focus required to work on complex problems.”
Mmmm. Yeah. I usually have to find a quiet place and eliminate distractions to get deep focus, but nice to know I can just carry this new device around with me and never lose deep focus!
The same as it's always been, which is to say you can live your life without worrying about it.
If our technology advances such that we can observe/find more and more of these, that doesn't affect the chances that a particularly sized asteroid hits the earth or not.
But I'm curious to know details about what this means:
> number more the 3 nodes in a PERT serial traversal, and do not have redundant options.
Would you mind elaborating or providing a bit more context I can go search on and learn about? Thanks!
I think at some point I'll need to turn it all off.
Man, those folks were impressive to be juggling all that information and have the communications processes in place manage it all... and calmly. (Steely-eyed missile men, indeed.)
I was born a month after they landed, and while growing up, the Apollo program was the inspiration for becoming an engineer.
> Who says you need to do networking in the kernel though?
I thought it implied they were bypassing the kernel entirely. This isn't true at all if it's still using a UDP socket to send... there's still a ton of code in the kernel to handle sending and receiving UDP. It's also a helluva lot less efficient in terms of CPU consumed, unless one is using sendmmsg, because it's a syscall per tiny datagram between user space and kernel space to get the data onto the UDP socket...
Just came to say I laughed and really appreciate the honesty in that statement. :)
(https://ratfactor.com/forth/the_programming_language_that_wr... is a great read, btw.)