|
oh lol it actually kinda existed, I thought it was some 'even number release version' thing
|
# ? Oct 18, 2023 22:29 |
|
|
# ? May 27, 2024 21:45 |
|
big v6 fan here, v6 is always the best
|
# ? Oct 18, 2023 22:35 |
|
ipv6 is a bit like python 3 in that they did a bunch of perplexing poo poo that can reasonably be described as "we don't actually want anyone to adopt this, even though we passionately say we do" ipv6 was standardized in 1995! We've been failing to adopt it for 27 years! what's that? DHCPv6 wasn't proposed until 8 years later, and was still a pile of debate and incompatibility for years after that? one of my favorites, RFC 4339 from 2006: quote:This document describes three different approaches for the lmao gently caress off
|
# ? Oct 18, 2023 22:52 |
|
the fact you can mix and match various portions of DHCPv6 (stateful or stateless) and SLAAC is one of the most baffling things about IPv6 autoconf
|
# ? Oct 18, 2023 23:05 |
|
rotor posted:so like what happened to ipv5? are we just not gonna talk about that at all? IPv5 and DirectX 4 eloped and their families disowned them
|
# ? Oct 18, 2023 23:35 |
|
here's my ipv6v2 proposal: it's ipv4, except for all addresses in 215.0.0.0/8 (the dod can just gently caress off to one of their other dozen blocks) the first 4 bytes of what used to be payload is actually just more address.
|
# ? Oct 19, 2023 07:04 |
|
just use numbers higher than 255?? doesnt seem hard imo
|
# ? Oct 19, 2023 07:57 |
|
need more space? 3462.15551.1394.7, done.
|
# ? Oct 19, 2023 07:58 |
|
it turns out that nat is just fine and actually preferable to a global address space
|
# ? Oct 19, 2023 08:07 |
|
Bloody posted:it turns out that nat is just fine and actually preferable to a global address space 99.99% of users won't notice being behind cgnat. the 0.01% that do will go "uh, hey, can I get an actual globally routable IP" and the customer support person just hits the "nerd detected, escalation required" button on their screen until the user gets a globally routable IP
|
# ? Oct 19, 2023 08:12 |
|
Bloody posted:it turns out that nat is just fine and actually preferable to a global address space how dare you, don’t you know i need every light bulb and appliance in my house to be a globally addressable host
|
# ? Oct 19, 2023 08:15 |
|
we'll have fusion power before we have global IPv6 adoption
|
# ? Oct 19, 2023 08:17 |
|
rotor posted:i always wonder what we'll look back on in 20 years and say "what the gently caress were we thinking" about https://forums.somethingawful.com/query.php?action=posthistory&userid=22993
|
# ? Oct 19, 2023 08:19 |
|
well-read undead posted:how dare you, don’t you know i need every light bulb and appliance in my house to be a globally addressable host there’s eight billion people on the planet and all of them can have little a cell phone, as a treat
|
# ? Oct 19, 2023 09:13 |
|
Kazinsal posted:99.99% of users won't notice being behind cgnat. the 0.01% that do will go "uh, hey, can I get an actual globally routable IP" and the customer support person just hits the "nerd detected, escalation required" button on their screen until the user gets a globally routable IP Don't modern console games often still require direct peer to peer connections for multiplayer, rather than routing through a central server? I found this completely baffling when I played Elden Ring, and when I had to set up port forwarding rules in pf.conf for Animal Crossing, I was wondering how normal people would ever manage this.
|
# ? Oct 19, 2023 09:57 |
|
rotor posted:i feel like just doubling it would be better. six octets sounds bad. eight octets sounds right. that should have been called ipv8
|
# ? Oct 19, 2023 10:01 |
|
Athas posted:Don't modern console games often still require direct peer to peer connections for multiplayer, rather than routing through a central server? I found this completely baffling when I played Elden Ring, and when I had to set up port forwarding rules in pf.conf for Animal Crossing, I was wondering how normal people would ever manage this. lmao
|
# ? Oct 19, 2023 10:03 |
i work in a mid-sized hospital as a blood scientist. primary care doctors just won't label samples with a collection time or stop storing them in the fridge overnight, which causes the cells to leak so much potassium we keep phoning out false alarms. being the most computer-touching person outside of actual IT, ive been asked to write a "if receipt at lab minus collection date/time > 24h, block potassium result" logic rule, to stop our staff wasting their time. today i went spelunking above the pathology IT staffs' desks, and as luck would have it, i found the exact combination of ring-bound manuals from the 90s and random photocopies from the mid 2000s i need to figure out how to make this ancient thing do what i want. whatever the gently caress UniBasic is, i have a manual now! healthcare IT is a gently caress, and the successor product we're forced to upgrade to next year is somehow worse than this ancient and venerable colossus.
|
|
# ? Oct 19, 2023 10:17 |
|
the only thing being worse than the old system being the new system is a normal state of affairs.
|
# ? Oct 19, 2023 10:38 |
Cybernetic Vermin posted:the only thing being worse than the old system being the new system is a normal state of affairs. the new system seems terribly designed (the logic engine works using visual basic, the patient search crashes the GUI if you run it on a too-common last name, with vendor advice being "please don't run big searches ", when i asked one of the vendors programmers if the software is 2037 compliant they said "i hope to be retired by then", and when i escaped their software window into the file system for shits and giggles, i found an installation log with an admin username and password printed into it in plaintext... that was a fun email trail with my boss, and a fun phone call from their head of compliance. at least our local project manager got a good laugh out of it). the vendors' employees don't help, either. on my last day at $prevEmployer, i had a teams meeting with the vendors' product specialist who'd been setting up our system. the system that decides whether we're alerted to "this patients' heart is about to stop lol" in time to do something about it. they were comparing the configuration files for the test environment and live environment. in microsoft excel, by pasting the plaintext documents next to each other and using a simple equals comparison to see which lines weren't exactly identical. i turned on screen sharing and showed them where i'd been pasting their config files into git for the past six months. "wow Lunar that looks really useful" yes. yes it is. they finished the call with "and what have you learned from this project, lunar?" "use source control?" "no, don't ask us for complex things" honestly, werk.
|
|
# ? Oct 19, 2023 10:54 |
|
Kazinsal posted:legacy x86 boot paths. CSM was supposed to be phased out by 2020, but every new board still has it, sometimes enabled by default. in 2040 you will likely still be able to boot a PC into a simulation of a 5150 leaving the past behind might happen sooner than you think we're now nearly 25 years since amd proposed x86-64, 20 since it became commercially available, everyone runs a 64-bit OS now, and 16-bit code is so far in the rearview mirror that anyone who cares can just run it in an emulator so, intel recently published a proposal for x86-s, a new version of x86 which removes the legacy bullshit nobody has needed for aeons. x86-s doesn't support anything but 64-bit mode (note, it can still run 32-bit userspace code since that's baked into x86-64 in a clean-for-x86 way, but the OS itself must be 64-bit). it also shitcans instruction prefixes which force 16-bit addressing even in 32- or 64-bit code, removes call gates, simplifies segmentation, axes virtual 8086 (rip 5150), and much more of course even if intel manages to get buy-in from microsoft and ships this anytime soon, i have no doubt that the pc industry will respond by baking an emulator for all the legacy stuff into the CSM, so yay! 5150 lives!
|
# ? Oct 19, 2023 12:38 |
|
Kazinsal posted:99.99% of users won't notice being behind cgnat. how do i know if im cgnat.?
|
# ? Oct 19, 2023 14:26 |
|
BobHoward posted:5150 lives! This is appropriate because the code for an involuntary psychiatric hold in California is 5150 and sometimes touching computers makes me feel like I need one
|
# ? Oct 19, 2023 14:28 |
|
Cybernetic Vermin posted:here's my ipv6v2 proposal: it's ipv4, except for all addresses in 215.0.0.0/8 (the dod can just gently caress off to one of their other dozen blocks) the first 4 bytes of what used to be payload is actually just more address. rotor posted:just use numbers higher than 255?? doesnt seem hard imo rotor posted:need more space? 3462.15551.1394.7, done. brilliant. now it's still incompatible with ipv4, but you now have no way to tell whether that ipv4-looking address is actually running on ipv6 or not
|
# ? Oct 19, 2023 14:48 |
|
BobHoward posted:leaving the past behind might happen sooner than you think i mean, about time, but the 16-bit support silicon on a modern x86 cpu core is a microscopic speck and it has no security implications so i wonder why they even bothered at this point. but yeah it would be cool. i think some embedded x86 parts do something like this already but i don't have a source for that.
|
# ? Oct 19, 2023 14:49 |
|
Athas posted:Don't modern console games often still require direct peer to peer connections for multiplayer, rather than routing through a central server? I found this completely baffling when I played Elden Ring, and when I had to set up port forwarding rules in pf.conf for Animal Crossing, I was wondering how normal people would ever manage this. no, its entirely up to the individual game dev. things made by developers who know what the internet is have servers hosting both pc and console clients. your two examples are: Elden Ring: a peer to peer multiplayer game from a company that doesnt really care about multiplayer Animal Crossing: a game on a platform from a company that doesnt acknowledge the existence of the internet
|
# ? Oct 19, 2023 15:07 |
|
And even when using p2p, a core, foundational feature of xbox live (and i assume playstation online services) is handling of NAT to make it transparent to the user. if your p2p or client/server networking doesnt work on xbox or ps its probably a bad developer. however if internet doesnt work on nintendo thats working as intended
|
# ? Oct 19, 2023 15:12 |
|
BobHoward posted:leaving the past behind might happen sooner than you think BobHoward posted:leaving the past behind might happen sooner than you think BobHoward posted:even if intel manages to get buy-in from microsoft BobHoward posted:microsoft holdin my breath here
|
# ? Oct 19, 2023 17:29 |
|
Lunar Suite posted:the vendors' employees don't help, either. on my last day at $prevEmployer, i had a teams meeting with the vendors' product specialist who'd been setting up our system. the system that decides whether we're alerted to "this patients' heart is about to stop lol" in time to do something about it. amazing
|
# ? Oct 19, 2023 17:31 |
|
NihilCredo posted:brilliant. now it's still incompatible with ipv4, but you now have no way to tell whether that ipv4-looking address is actually running on ipv6 or not in one case it is a dumb hack solution for which your complaint is obviously untrue, the other is a joke for which your complaint makes no sense (just look at the numbers, if they big it ipv6v2)
|
# ? Oct 19, 2023 17:32 |
|
imagine thinking it'd be hard to spot bytes with more than 8 bits smdh
|
# ? Oct 19, 2023 17:33 |
|
quote:Why does mojo have the 🔥 file extension?
|
# ? Oct 19, 2023 18:03 |
|
if you want to understand the modern NAT/UPnP/peer-to-peer landscape, soak up the dulcet tones of David Anderson: https://tailscale.com/blog/how-nat-traversal-works/
|
# ? Oct 19, 2023 18:20 |
|
whoever proposed this file extension should be forced to change their legal name to "💩 🏩"
|
# ? Oct 19, 2023 18:42 |
|
Cybernetic Vermin posted:in one case it is a dumb hack solution for which your complaint is obviously untrue, the other is a joke for which your complaint makes no sense (just look at the numbers, if they big it ipv6v2) I've seen that joke suggested in earnestness in the past (in the cosmetically different form of adding more byte blocks instead of widening the existing four) regardless of the encoding used, there are two scenarios: either an ipv6v2 device is allowed to use an ipv4 address as its sole address, or it is not in the former case, you get compat issues cause if you're 1000.1000.1000.1000 and want to talk to 6.6.6.6, you don't know if the latter is actually ipv6v2-aware until it tries to talk back to you and fails (and ofc you won't know if it failed because it's an old ipv4 device or for any other reason) in the latter case, where there is no overlap between ipv4 and ipv6v2 and all plain 32-bit addresses are reserved for old ipv4 devices, it's equivalent to the current ipv6 impl except your boomer coworker who never liked that newfangled dhcp thing can still yell the ipv6v2 address on the phone
|
# ? Oct 19, 2023 18:42 |
|
Internet Janitor posted:whoever proposed this file extension should be forced to change their legal name to "💩 🏩" https://en.wikipedia.org/wiki/Chris_Lattner
|
# ? Oct 19, 2023 18:44 |
|
chris lattner peaked with llvm and it's been all downhill from there
|
# ? Oct 19, 2023 18:45 |
|
DELETE CASCADE posted:chris lattner peaked with llvm and it's been all downhill from there i think you mean 💩 🏩 peaked with llvm
|
# ? Oct 19, 2023 18:46 |
|
JawnV6 posted:lol I mean. it’s happened before for amd64. and windows has been built for different cpus for a long time now
|
# ? Oct 19, 2023 18:50 |
|
|
# ? May 27, 2024 21:45 |
|
it's not the first time intel's proposed dropping legacy modes, and the last time they spat out a chip that did that, it was a super specific cut down embedded version of the 386 it got so little sales traction that they didn't do it again
|
# ? Oct 19, 2023 18:58 |