1,252 karma · joined January 28, 2020
It did a lot of performance wankery, like templating system (not my lib, I just used available one) being just Go code embedded in HTML that needed to be compiled with the rest of the app, or pre-generating HTML from Markdown on load. Using same language to write app as to write templates was nice tho.
Most of it was mostly "how far I can go without cutting features" because really going from 10ms per page to 0.2 ms per page has no difference after client RTT is involved.
Hell, even on localhost for some reasons Chrome always have few ms delay before starting download compared to any cli client, chrome shows anywhere between 3 and 8ms to start downloading, while FF sits at 0ms
I originally planned to have some fun with HTTP2.0 PUSH but, well, while I kinda believed on authority that it is useful for something it turned out to be entirely stupid idea which apparently nobody bothered to test before pushing it into standard so I didn't get to do it. Maybe I should try again with HTTP3.0 hints.
I tried various stuff but the common advice of "5 small meals across the day" just made me feel like some cow that constantly have to eat something and it was bother to manage, all while I was getting hungry quickly. Took a long while to switch but if I'm not trying to drop weight but "just" keep it steady I can eat big, nice, tasty meal once a day then don't touch food for next 24 hours without much hunger bothering me.
I stopped with caffeine because I was drinking it only at work and I often had headaches on Sunday, till I eventually figured out I will get headache from too much coffeine exactly 2 days after i stopped drinking it.
Didn't even drank it for the stimulation, just that I like coffee and work had expresso machine and I didn't... took like 2 weeks for me to stop being sleepy in the morning but wasn't that hard.
But they are designed for ESD protection, so while high voltage, not that much energy
The problem is that there is no fuses, while ethernet basically have transformer working as a fuse, USBkiller can hammer the protection diodes till they burn out and then the rest of it gets it.
> So running DC would probably just melt the cable or transformers but is not really likely to damage the device side (? Not actually sure about this).
A step of voltage, whether that's just turning on AC on top of the cycle or just pushing DC thru it would go thru the transformer in form of short spike.
So sure, looking at random transformer data they are protected to like 1000Vrms so it wouldn't jump thru isolation, but initial magnetic spike might, and then it is up to any surge protection on the other side to ground that
> But in this case it runs mains AC.. so... Do we expect the cable to blow first due to the high current, saving the device, or do would we expect the brief period of AC would cause enough induced current on the device side to also fry that?
I'd guess transformer would be the highest resistance/area thing in the chain so it would melt first.
Or how in some hydraulics system it's not really the flow that's the motivator for action, like in simple hydraulic brakes it's not the flow of the fluid in a circle that powers anything, it's just that the push on fluid on one side pushes the brake pad on the other. So it's not really a flow but wave of pressure (which is kinda how current actually flows in the conductor anyway...)
SPOE/SPOA added a bit of programmability to HAProxy but it is still basically only messing around with headers and acting upon that, nothing to do with content.
That's actually how the big generators are synced, if there is less load you will see frequency of network increase slightly until the amount of generation drops, similarly in other way.
Inverter based one, hard to tell, entirely depending on code driving it.
But taking into consideration even 20V over low resistance wire could be 20+ amps it would probably trip pretty quickly.
> California already forces people to spend $1000's on LED compatible dimmers and moisture / motion sensing switches (spoiler alert: none of these things work properly, and they all have a high first year failure rate). They also have fascination with putting power outlets in useless places.
"This stupid thing for $100 doesn't matter, we already have stupid thing for $1000" is not a great argument either.
> I even burned them many times, to the dismay of my dad. I burned them because I wanted hotter water, and closed the water line as much as I can, the resistance heated up to red and then melted.
I'd imagine making one with PTC (element that decreases resistance when heated, self-regulating temperature) would prevent that but that would add a buck or two on material bill
Still can't watch him, I get anxious
> instead of three: (female-male, female-female, male-male).
One. Male to female. From where power flows to where power is delivered.
There is no reason for F-F or M-M power cables to exist.
The whole point is to have power source never be male as that's exposed, energised metal.
People needing that cable are just being cheap and not wanting to do it properly
you'd have something like
gen -> circuit A -> main bus -> circuit b-> load
|
main fuse/switch
|
outside
So loads on circuit A would not be protected by any fuse but ones at circuit B would.Also circuit A would allow full generator amps to the devices there, no RCD protection either.
I dunno, statistic about deaths from electrocution?
There is Bareos/Bacula but that just pretends everything is a tape and generally works badly/quirky because of that.
* have a bad block on one drive * have a bad block in *EXACTLY SAME LOCATION* on the second drive * have a bad block in *EXACTLY SAME LOCATION* on the third drive * all of that happening within one scrub period (which is usually once a month).
And that's just simple RAID6 setup without external redundancy.
As you can see it would be far easier to lose data due to just correlated drive failures, than from random block corruption.
Starts to matter where your k8s address can be <service>.<namespace>.svc.kube.example.com
> Merely looking at the options for my DSL device, my god man, even after all these years in tech, I'm still a bit bewildered.
A lot of DSL dumbfuckery came from telco. For example, ADSL uses ATM framing which is 64 bytes, with 48 bytes of payload. And it is fixed size packet which means mere TCP ACK, with all header overheads, won't fit and you need to send at least 2 ATM packets for each TCP ACK AND waste some bytes in them too. ATM was first used for voice which is where constant size came from.
The VPI/VCI settings are also directly related to ATM protocol
And why 48/64 byte ? Coz EU wanted 32 byte payload (coz they wanted lower delays) , US wanted 64 byte (coz they wanted lower overhead) and they just split the difference in the middle.
Why the fuck they applied it to inherently point to point ADSL ? Who the fuck knows, you'd think throwing away 25% bandwidth on ATM header would be bad for residential service over shitty wires but they apparently didn't care.