Hacker Newsnew | past | comments | ask | show | jobs | submit | pa7ch's commentslogin

Risk limiting audits are why this work. You physically sample ballots at random. The number you sample grows as the gap in the electronic tally shrinks to reach high confidence the election was tabulated correctly.

The normal person has no knowledge of stats. I am a professional physicist, and I struggle with stats. The methods you suggest can convince a stats professional that the tally is correct. It cannot convince a normal person of the same.

> It cannot convince a normal person of the same.

But you don't need everyone to be convinced of it first-hand. You just need everyone to trust someone who is convinced of it.


Election security should not hinge on a "trust me bro" - especially when people are being convinced the other way by Russian propaganda talking heads on social media.

Manual counting requires zero trust. In my country anyone is welcome to observe the entire process from start to finish, if they wish to do so. A few years back a fringe far-right party tried importing the voting integrity distrust over here, and recruited people to watch their local polling stations to "expose the fraud". Which was totally fine because they were always allowed to do so, and it fizzled out because zero evidence of fraud was found, and that party still didn't get a significant number of votes.


Don't these two situations (watching vote counts; understanding a complicated statistical argument that the vote is tamper-free) require the same kind of trust?

1. In both cases, everyone is theoretically capable of checking it themselves; they're theoretically zero-trust. In the former scenario, I'm theoretically capable of attending the vote count, and in the latter scenario, I'm theoretically capable of learning the statistics needed to verify the argument.

2. In both cases, most people cannot (or will not) practically check it themselves, and is trusting that someone they trust is doing the checking for them.

I'm not saying they're the exact same situation, but they both ask for a large amount of trust from most of the voters.


You are correct on both points, which you elucidated well. Let's me differentiate the two systems based on "who-to-trust".

- The observe-system operates on an adversarial basis. The people observing the voting process are state officer, independent observer, each party's observers. If you vote for party X, then you trust that party and its people to do right by you. This include trust party X's observer, who additionally is often a local well-known person. You can actively distrust all the other observers and officers, and as long as your observer gives the A-Okay, you are happy with the result. This trust in your observer is a very simple human kind of trust. No expertise is needed by your observer. If you trust other observers, your trust in the result goes even higher.

- The stats-system is founding its trust in the trustworthiness of the stats experts. The problem is that (1) you don't know the stats expert personally. In fact, a huge chunk of the population in any country doesn't know anyone who is good enough at maths and stats. If people in your family are not the math type, your friends will also not be the math type. (2) It is incredibly easy to sling mud at the expertise and trustworthiness of an expert. This process is operating at a very high level these days on social media. Anyone remotely connected to politics is continuously character assassinated by others. Adopting a stats-system actually will actually increase this mud slinging to new heights.

The observe-system is better because as someone else has said, all the counters and anti-counters to it have been known for 100s of years. Breaking it requires breking 100s or 1000s of polling stations across the country. The stats-system has more central points of trust which can be broken more easily.


He has been living in LA and working for the EFF for some time now.


whats wrong with usb-a? I feels more sturdy and less likely to have connection issues then usb-c in my experience.


> whats wrong with usb-a?

Which way up it should go.


Simple. The third way you try, always.


Macro-scale quantum hardware: USB-A is a spin-1.5 connector.


The side with the holes. That's true for 95% of devices, with one of the few major exceptions being cheap chinese powerbanks


Or if you’re plugging into a desktop where the ports are vertical


There is usually still a concept of front. Most desktops if laid down would be laying on their right side (so the motherboard & cpu aren't upside down). From there you can still pretty easily tell how the port is oriented without looking.


Which requires being able to see the plug.


As with most plugs that came before it.


They were not externally symmetric


Also the side that has the logo on.


USB-C is better than A in that it works in two orientations instead of one, but the correct answer for connectors should be any orientation — the best connectors are cylindrical connectors: barrel plugs, RCA, BNC, banana, phono, TRS, TRRS, etc. Just make them round.


Man, there’s nothing more satisfying than the feel of a quarter-inch TRS plug slotting in to a high quality jack. Truly one of the great plug designs.


And one of the oldest. It dates back to the 1870s. Older even than the Edison Screw.


LEMO push-pull connectors and their ilk are mighty satisfying.


Well, more like 1.5 orientations because we STILL have devices that only work with the cable in one of two valid orientations.


Would it be practical to have a round port as a universal connector? USB C uses a lot of pins, how would that work? Like an audio plug with a lot of rings?


I think it would be practical with glass fibre. Two wires/rings for power, and fibre for data. Something like a Mini-TOSLINK, but even smaller. Ideally the plug would be barely thicker than the cable.


Glass fibre is pretty fragile though, it would probably break in the first hour for most people if used like a normal USB Cable.


Just use polymer fibre for consumer gear. Cheaper and more robust.


Can that still transfer data at full speed with reasonable cable length while bent?


The other way


No, the other other way.


It's almost impressive that they designed a port that feels so wrong when you actually get it right


I worked in an IT department at one time and encountered USB-A plugs forced into Ethernet ports.

It seems so unlikely that I’ve just searched it to see if it’s possible, but am getting no hits.


My laptop has one of these ethernet ports that half close when not in use. It doesn't work anymore because someone mistook it for the USB port that's right next to it when distractingly plugging their keyboard in.


This reminds me so much of the old story about how a user called support to tell them their floppy drive wasn't working. When the tech got there, the computer had no floppy drives, and the user had been forcing the disks in the gap between the drive blank plates.


no, they definitely fit. They're just awkwardly exactly the right size that while you're trying to plug things in punched over under the desk and crawling around and feeling around the backside; it just yeah.


PS/2, which USB all but replaced, solved this by visually keying one side of the connector as flat.


And I have more than once bent the tiny pins by trying to orient the plug by feel, and it was a nightmare to fix it afterwards.


You've never ruined a PS/2 port by rotating the plug to find the right orientation?


When I'm trying to plug my PS/2 keyboard into the port in back of my computer which I cannot see, instead of needing to try two orientations, I need to try every orientation.


And USB solved it by having the logo on the upper side.


Where the logo is.


And when the port is vertical and you can’t see it?

I’m surprised how tolerable people seemed to find Apples rear ports.


It's very weird that USB-C solved the problem of "we can't tell which way to insert the plug" by mandating that both orientations should work, as opposed to just making the exterior of the plug as asymmetrical as the interior.


I don't find it weird. Not even having to work out a correct orientation is a great convenience. The micro-USB connection (or is it "min"?), which I need to fiddle with to charge some older gadgets, is a testament to how annoying an "asymetric exterior" plug can still be.


With micro USB you end up with damaged plugs and ports in my experience.


Yes, micro USB is far too flimsy for a lot of things it’s used for from what I’ve observed. The connector seems to have a lot of leverage for ripping its tracks off, but often not a great mechanical connection to the board.


You mean something like HDMI? If you’ve ever tried to plug one of those into the back of a TV, you’ll know it’s still pretty difficult to get it the right way up.


> If you’ve ever tried to plug one of those into the back of a TV, you’ll know it’s still pretty difficult to get it the right way up.

That's true, but the difficulty in that case comes from being unable to see the hole or fit into the space between the television and the wall.

For example, plugging an HDMI cable into the back of a monitor involves none of the difficulty of plugging an HDMI cable into the back of a TV, even though the connector and the port are the same in both cases.


At least HDMI is a 'low frequency' connector, often only ever plugged in once, as opposed to USB (or refueling a car)


I bought a cheap USB hub so I don't have to reach behind the TV to plug things in


Have you ever tried to use a D-sub connector?

Very asymmetric, impossible to plug in backwards, only one possible orientation.

Still absolutely fucking awful. Same for HDMI, displayport.

The RJ series are the only asymmetric connectors I've use that you can easily get in blind.


Less weird as they get smaller. Call it an accessibility thing if you like, but I think it's better for everyone and congrats to them. Isn't this what technology is supposed to do, make things easier?


sometimes you're plugging in things at the back of something nearly flush against a wall and you can't really see, its quite useful for the connector to be reversible.


Very weird to market this as subscribing to "Elon process for software"

I remember when defcon ctf would play Geohot's PlayStation rap video every year on the wall.


I hate it when ‘inspirational’ quotes are attributed to the person with the largest audience and not the people who came up with it, like in this case, the engineers at Lockheed’s Skunk Works.


It's an apocryphal quote.

"A designer knows he has achieved perfection not when there is nothing left to add, but when there is nothing left to take away."

- Antoine de Saint-Exupéry


Or Colin Chapman at Lotus: "Simplify, then add lightness." Could just be wise advice :)


I like this. A lot.



Ok this wins my “best wikipedia article of December” award


Check out this cacophony of an infographic:

https://upload.wikimedia.org/wikipedia/commons/b/bf/40_princ...

No wonder the soviet union collapsed!

When you need Moscow to explain the hammer sign on your corded drill, its too late for damage control.


Honestly seems like zig is shaping up to be a better fit for kernel. Regardless the language that attracts skilled kernel devs will matter more then lang.


Rust has been stable for 10 years. Zig is still not stable.


It is not - and given its stance on memory safety, it will hardly ever be.


It's true only for unsafe rust.


Gokrazy is a minimal linux distro that just boots into a go init program. You can run on a raspberry pi or pc. It has a little init system that just takes a path you normally use in `go run` and just runs them and restarts as needed. Its been a joy for me to play around with. Has A/B updates as well.

https://gokrazy.org/


I feel like wc3 is undersung. To me it achieved the perfect balance of allowing potentially mechanically worse players to win with brilliant tactics or strategy. It put the emphasis on strategy in rts more then anything else.

As a kid I was shit at it and played customs maps and goofed with the editor. Now I've gone back to find grubby streaming and revealing the depths of the meta evolution, and counters.

I like that even when a strong meta develops people can potentially counter with strategies that aren't as well rounded for long term use but upset the current meta.


I think its fundamentally more difficult to host communications services where spam is possible and there is no auth/contact system in place before first communication can happen.


I'm not an expert in this area but from what I understand what was once novel content spam filtering is not at all novel now, there are easily trainable model strategies (BERT?) that get you to 99%.

A whitelist, auth/contact is ideal for messaging without spam and is more workable with a large federated group that can adopt an evolving open source protocol.


I was a big fan of the original check handle proposal: https://go.googlesource.com/proposal/+/master/design/go2draf...

I see the desire to avoid mucking with control flow so much but something about check/handle just seemed so elegant to me in semi-complex error flows. I might be the only one who would have preferred that over accepting generics.

I can't remember at this point because there were so many similar proposals but I think there was a further iteration of check/handle that I liked better possibly but i'm obviously not invested anymore.


Its the opposite in my understanding. Zig has a BDFL.

Trademarks are the usual cudgel of choice to enforce a bad actor claiming to be part of offcial Zig.


But he isn’t. He’s just writing an AI slop book about Zig. Surely there’s nothing legally wrong with that? He never said it’s an official book or backed by the Zig project.

The trademark cudgel is used on people who release an incompatible language that they insist on calling Zig, confusing people who want to try Zig. Or people who add malware to the Zig tool chain and try to distribute that.

Trademark can’t be used to control bad actors like zigbook.


> Surely there’s nothing legally wrong with that?

Incorrect. Not honoring the attribution requirement in the MIT license is a copyright infringement because it violates the terms of the license, which are legally enforceable conditions.


We are specifically talking about what the Zig project/foundation headed by Andy Kelley can do to such bad actors using the Zig trademark - which is exactly nothing.

I wouldn't be so quick with the "incorrect" if I were you. You haven't even taken the trouble to read two sentences.


> I wouldn't be so quick with the "incorrect" if I were you. You haven't even taken the trouble to read two sentences.

I wouldn't be so quick with the dismissal if I were you. You haven't even taken the trouble to read the article.

Also, Quad erat demonstrandum - the infringing repo no longer exists.


Mm thats a good point. I'm not entirely clear on the limits of trademarks in this case. Its Zigbook rather then Zig.


I read a lot about this when Rust was considering adopting a trademark policy. The main use cases for enforcing the trademark were

- preventing someone who hardforked the project from creating an incompatible language while using the same name.

- preventing someone from distributing malware while still using the same name.

Because if you notice, neither of these clash with the MIT license that many languages use. You need to enforce your trademark to stop this kind of behaviour.

Zigbook can argue that they aren’t causing any confusion between themselves and the Zig language. The Zig foundation could argue that the name implies an endorsement by the project and they should call themselves The Unofficial Zig Book instead. I don’t know which way it goes.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: