Mats Järlström’s victorious 6-year battle over yellow lights | Hacker News Reader

Mats Järlström’s victorious 6-year battle over yellow lights

koin.com

180 pointsGarbage133 comments

Quick summary: when someone is making a right-turn, and the light has changed to yellow, they do not have sufficient time to finish their turn before the light turns red. Therefore, at lights with red-light cameras, many people who could not possibly stop in time are getting tickets. (The point is that the fact that cars turning have to slow down must be taken into account when calculating how long it takes to traverse the intersection. If you are going at close to the speed limit, you get through while yellow, but if you have slowed down because you plan to turn, and the yellow occurs too late for you to stop, you will be long enough in the intersection to get a ticket).

Mr Jarlstrom made many efforts to point this out to the Beaverton (OR) city council. Not only did the officials treat him with disdain, the state even fined him for practising engineering without a license. (Talk about shoot the messenger!)

However, it turns out he was right, and he has been vindicated in an article in the Journal of the Institute of Transportation Engineers.

Fined for practicing engineering without a liocense. Institutions will always protect themselves but this is absurd

Not really. I mean, fining this person was absurd. But licensing the title "engineer" isn't absurd; bad engineering gets people killed.

> fining this person was absurd. But licensing the title "engineer" isn't absurd

You can't logically have one without the other. If he claimed to be an engineer (which he acknowledges) and the state claims the title for licensing (which it did) then you can't enforce that licensing without some sort of penalty.

You easily can, by being more strict about what constitutes the use of the title of engineer.

Exactly: sign off on civil works plans: licensed. Charge for advice related to:: licensed. Everything else: not.

Sure you can. You can make a legal requirement that somebody has to be licensed to sign off on things that require the license without fining for the use of the word. Imagine that you're an engineer in Germany, you go to Oregon and somebody asks you what you do for a living. You reply with "I'm an engineer" and get fined because you're not licensed in Oregon.

In an oddly mirrored version, up until about 12 years ago it was illegal for someone with a PhD from outside the EU to use the title "Dr." in Germany, without going through specific paperwork to get permission to use that title.

http://htor.inf.ethz.ch/blog/index.php/2008/11/06/you-can-al... quotes a Washington Post piece:

> At least seven U.S. citizens working as researchers in Germany have faced criminal probes in recent months for using the title “Dr.” on their business cards, Web sites and resumes. They all hold doctoral degrees from elite universities back home

A NZ PhD holder in Germany could get in legal trouble until at 2012 - http://www.stuff.co.nz/editors-picks/6783089/Germany-goes-fo... .

I really don't understand this backing of the state here.

There is a vast gulf between stating one is an engineer, which many people are, and that one is a PE.

Reminds me of zero tolerance rules at schools, where adults claim that they can't understand the difference between two very different things.

> I really don't understand this backing of the state here.

You're just letting your pet peeves color your perceptions. I'm not backing anything here, I'm pointing out that if you're licensing the title of engineer then it makes sense to protect that license and therefore penalise people claiming the title without being licensed.

> There is a vast gulf between stating one is an engineer […] and that one is a PE.

In Oregon, before this lawsuit, there was not.

> Reminds me of zero tolerance rules at schools, where adults claim that they can't understand the difference between two very different things.

Yes, your comment also reminds me of people refusing to understand basic concepts.

The simple point was that Oregon's position was foolish. You keep pointing to Oregon's position as if it makes any sense. When one sees a bad government policy used improperly you don't usually state "But the state says so, so I guess it's just the way it is".

It's a bad position and this case is an example of why it's a bad position.

You stated that "If he claimed to be an engineer (which he acknowledges) and the state claims the title for licensing (which it did) then you can't enforce that licensing without some sort of penalty."

Of course you can. He didn't build bridges for pay, which is the type of engineering the law is clearly about. He just stated that he was an "engineer".

But yes, I'm the simplistic thinker, for pointing out that shades of gray exist in your black and white definition of things.

What about software engineers?

The simplest way to resolve that dilemma is to acknowledge that almost nothing we do is "engineering".

You can also overstate how much engineering generally is about rigorous processes and theoretical correctness as opposed to heuristics and empiricism.

I don't actually disagree with your general point. But there are plenty of examples of civil engineering project failures because of defective materials and the like and there are established practices in many areas of software.

I've worked in engineering outside of software--and was even on track to get a PE--and a lot of that was pretty ad hoc.

I'm still kind of haunted by this very concise and blunt argument by 'jcranmer:

https://news.ycombinator.com/item?id=22315607

(I hadn't read about the KC Hyatt disaster before).

The Hyatt accident was a pretty massive screwup. There are other examples. See, e.g. http://www.slate.com/blogs/the_eye/2014/04/17/the_citicorp_t...

OTOH, Heartbleed had as much to do with critical open source code being maintained by someone who was basically doing on a shoestring via donations as a lack of software engineering processes in general.

It's not so much Heartbleed; I agree, that was kind of sui generis. It's just the more general sense in which our field has no guardrails to prevent people from opting for faster/cheaper time to market at the expense of security and reliability. Everyone in this industry is constantly drilling holes through the support beams and hanging whole new floors off them; the buildings collapse every week, and we just shrug.

I'm not even saying things must necessarily change. I'm just making the case that what we're doing isn't engineering.

A proper train wreck requires an engineer.

Proposing, designing, building, maintaining, scaling, duplicating, etc complex systems that handle billions of dollars in commerce a year, billions of events a day, petabytes of data, etc.

Successfully doing that is a lot, lot more than being just a software "developer".

This is an intellectually shallow dismissal and it frustrates me when I hear it made. The argument is most often used to diminish the status of software engineers; it has virtually no other truth value, but if we actually accept it, it does have consequences.

Planes don't stay in the sky, money didn't arrive in a bank account, technocratic administration of the government didn't continue another day, email didn't continue to hit inboxes, phones didn't continue to work, etc etc etc, just because amateurs got improbably lucky again today.

Our field has its challenges, but given the number of people who didn't die today because it succeeds the overwhelming majority of the time, it doesn't make sense to diminish the people addressing those challenges by deprofessionalizing them. Additionally, the practical impact of deprofessionalization would be "Software can't be trusted to the geeks. Let's give it to competent professionals, like the executive branch of the government. We have replete evidence they'll do a better job, mostly via writing requirements documents to give to Raytheon."

A quickly following second order effect of licensing engineers will be an extremely predictable one-two punch of a) differing career paths for licensed engineers and commodity code-monkeys and b) demographic differences in entrants into those two pools along axes that you will probably care about.

All it takes is one tick box requirement in federal contracting templates ("Vendor certifies that all engineers working on this system and subcomponents are duly licensed") to infect AppAmaGooBookSoft's HR policies. AppAmaGooBookSoft will institutionally love that requirement; people who care about the future composition of the industry's hiring pool should not.

We'll get to exchange being plausibly the most accessible white collar profession in the US for one which will be aggressively and formally policed at the point of entry by the same sort of folks who convince e.g. the ACM to advocate for immigration restrictions. Many folks are annoyed at the extent to which individual engineers engage in gatekeeping behavior. Few would consciously choose to advocate adding professional gatekeepers who have quarterly targets for gatekeeping productivity. That's not an incidental effect of a licensing regime. It is what licensing means.

I know you don't particularly like CATO, but CATO will win the argument about what will happen under a licensing regime by an overwhelming margin and, even taking into account you and CATO have very different values, you would hate the downstream impacts of that policy.

(This is perhaps a lot of effort for a lowbrow dismissal that was probably more intended to vent than be a policy proposal but this particular one is a 20 year hobbyhorse of mine. Our desired regulatory stance as a profession is not consequence-free; if we'd actually want a licensing regime, it should be because we've done the math and are willing to accept the consequences. If not, we should loudly and specifically reject the characterization that we're not equivalent to licensed professionals.)

I don't think you've actually responded to my argument, or any argument I would be likely to make. I am not in favor of licensure for software developers and feel like I've said as much on many occasions.

I'd like to give you a dozen paragraphs in response to your rebuttal, but unfortunately I don't need them to refute you. All I have to say is "none of this makes what we do engineering".

We are not engineers. We have none of the guardrails serious engineering has. We normally make decisions in an almost complete absence of consideration for safety, reliability, and security. Buildings erected millennia ago stand today, doing what buildings generally do, and even throughout the worst of the industrial disasters of the late 19th and early 20th centuries, buildings generally managed to keep themselves upright. Nobody suggests that civil and mechanical and structural engineering haven't progressed since then, didn't address vital safety problems, and didn't formalize their respective disciplines. They obviously did. But that's essentially the claim you're making about software. Just because planes don't routinely fall out of the sky doesn't mean we're not in the Triangle Shirtwaist era of software development

Bad engineering shouldn't be related to use of the general English word, nor should it be used as justification over ownership of such a word. I admit I am unfamiliar with this specific case, but if he claimed he was an "engineer as certified by X" I'd be more sympathetic to X claiming he's lying.

It's the same as saying you're a lawyer.

In the US, if I say I'm a lawyer, I'm at least implying that I passed the bar in some state (although possibly not in your state). If I just went to law school or if I just read a few books on constitutional law, it wouldn't be normal to describe yourself as a lawyer and you wouldn't be legally allowed to represent yourself as one.

People who have even grad degrees in some branch of engineering and have been working in the field for decades. But simply have never had a reason to get a PE? Perfectly reasonable for them to describe themselves as engineers (but not, of course, as PEs).

More like saying you're a teacher in this case. While using that title to get around regulations teaching others in certain settings could be a crime, use of the word colloquially while generally giving knowledge to others (especially if you were a teacher where you came from) should never be a crime.

In this case, yes. But more generally --- the context of the top comment on this subthread --- no.

Except going slower means you can stop faster, and considerably more so than the linear progression of time. So really the undue burden must be on people going straight through the intersection at the speed limit.

From the journal article:

"GHM’s solution to regulate a yellow change interval first appeared in the 1965 ITE Traffic Engineering Handbook, and it has become known as the kinematic equation. However, GHM’s solution is limited to vehicles traveling through level intersections at constant velocity, which does not include vehicle deceleration to execute safe turning maneuvers. This article presents a brief review covering GHM’s original solution and Mats Järlström’s extended kinematic equation which allows for vehicle deceleration and turning maneuvers."

By the time the yellow comes up, you've already made the go/no go decision, and you are scanning for pedestrians and other cars. You don't even see the yellow.

We had this problem in Slovenia — cars chasing yellow and thus creating unsafe intersections.

Solution? Yellow means stop. If you run a yellow it’s the same as running a red.

You can only run a yellow if you’re already in the intersection when the light turns.

Despite the groaning and complaints from motorists, pedestrian impacts and high speed intersection collisions have dropped dramatically.

This doesn't work though since you can't instantaneously stop.

Imagine you're going 20mph/30kmh and the light turns yellow when you're 1 meter away from entering the intersection. If you try to stop, you'll end up stopped in the middle of the intersection.

If your response is "in that case just keep going to clear the intersection"... That's exactly what the existing yellow light law is here that is being debated in the OP.

> If your response is "in that case just keep going to clear the intersection"... That's exactly what the existing yellow light law is here that is being debated in the OP.

That is my response indeed.

The problem arises when you treat yellows as "clear the intersection" and you get people stepping on the gas who are 100m from the intersection. This often leads to high speed collisions and pedestrian fatalities because you now have a car that instead of just rolling through a red, is absolutely blasting through.

Or at least blasting through a yellow.

I see this scenario in SF all the time. Car driving 25mph, sees light turn yellow, speeds up to 50mph to make the light. Could've stopped in time if they treated yellow as "stop".

This is obviously dangerous, but totally legal under "you can ride through yellow" laws as the car did in fact clear the intersection before the light turned red.

Speeding is still illegal

The yellow light absolutely needs to give sufficient time to clear the intersection. If it doesn't, that's a problem in itself.

This is dumb as why even bother with yellow.

To give cars within stopping distance time to stop, but discourage them from accelerating to beat the red.

No, seriously, how does that differ from red?

If you're in stopping distance and the light changes to STOP, then you STOP, no matter whether stop is encoded as red or yellow or both. So you don't need yellow for that, you can just give them red. They will stop.

If you're not in stopping distance and you get a sudden STOP sign, whether that's red or yellow, well you're kinda fucked because you're going to end up in the intersection anyway.

I don't see any rationale for having yellow with these rules (except to indicate that the light is about to turn from red to green).

The issue is that you’re in stopping distance and light changes to “Soon STOP”, so many people accelerate to make it through before the light says “STOP”

If yellow means “STOP but grace period if you cant” it removes the incentive to accelerate

The fact is that you must not run enter the intersection on a red light. If a green changed to red abruptly, it would be impossible not to do that.

The purpose of yellow, with a decent interval, is to make it possible to stop on a red, provided a vehicle is in good condition and not egregiously speeding, and the driver is sober and alert

Word semantic games about how that is expressed in the language of the traffic code do not change how it works.

"STOP but grace if you can't" wording makes it appear as if under some circumstances, it is a violation to enter the intersection on yellow. It creates incentive for police to bust people who go through yellow lights, on the premise that the judgment of what is an acceptable way to proceed through a yellow light is some objective standard that can be accurately policed.

I got a yellow light ticket by a cop who went through the same light, some 15-20 meters behind me! I remember thinking, "I was totally okay, but that driver behind me was pretty late; was it still even yellow?" Then that car put on police flashers.

I took it to court and prepared the defense that, if a vehicle A has time to stop on the yellow, then any vehicle B that is behind it must have even more time to stop. If the driver of B judges that he has no time to stop and proceeds, but claims that A ahead of him had plenty of time to stop, that is an inconsistent, double standard. (Actually, that was my back-up argument; my main argument was that I didn't want to be rear-ended by the car behind me, which clearly intended to proceed. And I would call as the witness the driver of that car who happened to be fortuitously present in the courtroom.)

I never got to present my argument, because that officer testified as having no evidence, so the case was dropped.

Anyway, sure, people will only change their behavior, but only due to being worried for being ticketed for going through a yellow in a manner that some officer doesn't like.

Here is a thought: if you are certain you do not have time to stop, and you're below the legal speed limit, why shouldn't you accelerate, if you feel like it? Say I'm doing 35 km/h in a 50 km/h zone (maybe I just left a parking spot, or turned a corner), and I'm 1 meter from the intersection which is green. It turns yellow. Why wouldn't I accelerate, or continue accelerating? Accelerating, by itself, doesn't prove that you had time to stop.

Well, the "why not just" approach would be:

- Traffic violation if you enter the intersection when the light is RED.

- Traffic violation if your speed increases at all when the light is YELLOW.

- Yellow light must last at least Ty = A * L + T where: L is the posted speed limit, A is a accelleration that lawmaker must be exposed to 100 1-5sec intervals of annually before renewing the law establishing values for A and T for that year, and T is a reaction time that lawmakers must consistently (ie 100/100 tests) be able to meet (also annually).

> You can only run a yellow if you’re already in the intersection when the light turns.

The problem is your previous statement implies that if you are going 45mph and the light turns yellow when you are 5 ft away from entering the intersection, you are required to stop. The physics doesn't work.

Where I live, as long as you're in the intersection before the light turns red, you're fine.

In Oregon, where this all took place, that is not the case.

The engineering association didn't really fine him for the campaign itself. They were upset that he claimed to be an engineer when making his arguments, despite not being licensed in that state. Even then, they gave him repeated warnings before issuing the fine.

In the end, the courts ruled in his favour. As a matter of freedom of expression, anyone can now claim to be an engineer in the state of Oregon.

Hah, the real world version of an internet argument I see time to time wherein people with engineering certifications are put off by software developers calling themselves engineers.

To be fair to Mats, I think he was a fully qualified engineer in Sweden. He was quite unreasonable in insisting on calling himself such in Oregon, but I suppose reasonable men don't change the world.

I didn't realize that one's knowledge and skill is a function of location.

Different locations have different engineering codes. This is in part why US states do not recognize each other's PE assignations.

I can understand why he couldn't be hired as a licensed engineer because of that, but that doesn't make him not an engineer. He wasn't arguing over something that relies on local standards anyway. A second is a second regardless where you are.

FWIW, there are different types of seconds. Eg, Universal time seconds based on the Earth's rotation, and atomic clock seconds based on consensus. And leap smear seconds are variable length seconds to help merge one into the other. https://developers.google.com/time/smear

I also find it intriguing that all of the Professional Engineers in Oregon, including ones involved in setting up the timing on the aforementioned yellow lights, failed (or chose not to) solve a basic kinematics problem like Mats did. Why did it take someone who was not a PE to fix this issue?

In some jurisdictions (like Canada), the title "engineer" is a legally-protected term. Someone like me who has 3 engineering degrees -- but have not taken any licensing exams -- cannot officially use the title "engineer" on a business card without being liable to a fine.

I can sort of see the intention behind laws like this, but I also think it's a bit of gatekeeping. The original intention was to prevent someone who isn't licensed from providing professional opinions and from signing plans without also being accountable/responsible for the outcomes.

This is all well and good, except there are many engineering disciplines (outside of civil, mechanical, electrical... and even within them) where signing plans and providing professional-grade opinions are not the norm. My father was a practicing engineer for 30 years and has never had to sign a single plan, and so has never taken a licensing exam.

A licensed engineer typically has an additional P.Eng. or P.E. title (Professional Engineer), and I agree that that title should be regulated like all other licensing titles. But the word "engineer" is so generic that it doesn't really make sense to try to legally protect it.

Are you sure you can't use the title "engineer"? I thought it was only the very specific term Professional Engineer that is legally protected, and what someone looking to hire a licensed engineer would be on the look out for.

Edit: According to wikipedia, there are plenty of titles in Canada like locomotive engineer used by a train operator even though the position is unrelated to the Professional Engineer license [1].

[1] https://en.wikipedia.org/wiki/Regulation_and_licensure_in_en...

Pretty sure, unless the law has changed from when I took the required professional practice courses in engineering school. (I went to schools in Quebec and Ontario)

That part you cited in wikipedia has no citations and may not be legally correct. This is something that is codified in Ontario law [1] (not sure about other provinces, but I believe it is harmonized federally -- in Quebec the title ingenieur/engineer is protected.). This law has been tested several times (example in [2]). More info here [3]. Even Microsoft had to give up the use of the term "engineer" in Canada after being hit with lawsuits [4].

[1] https://www.ontario.ca/laws/statute/90p28#BK43

[2] https://www.peo.on.ca/engineering-licensing-body-clarifies-u...

[3] https://engineerscanada.ca/frequently-asked-questions

[4] https://www.canadianconsultingengineer.com/engineering/micro...

That doesn't explain how locomotive engineers are allowed to keep engineer in their title? Given 1 literally states ` uses the title “engineer” or an abbreviation of that title in a manner that will lead to the belief that the person may engage in the practice of professional engineering;` it all seems like bullying from the org trying to lay claim to Engineer so they can collect more dues when "Professional Engineer" vs "Engineer" is enough of a distinction.

also 4 is directly relating to training certs for a "systems engineer" not job titles which does make some sense, microsoft still seems to use the title "software engineer" in canada for jobs

(1-3 are not what i would consider unbiased sources)

I don't know enough about the legal nuances to discuss or to challenge the PEO, but these are the facts as I know them.

In the context of what was basically "I have been educated as an engineer, which is why you should be giving me more credence than a random person", I feel the use of the word is perfectly fine.

In the real world, professional engineers are bonded and are liable for errors like lawyers.

Are there many examples in the real world of this actually happening?

Close to home... yes, the FIU bridge engineers could have been charged with manslaughter (https://www.miamiherald.com/news/local/crime/article23659748...)

I'm pretty sure anyone that works on something that kills people "could" be charged manslaughter. The entire PE system seems antiquated and obsolute.

Yes but that has nothing to do with the State Board of Engineering Examiners or whatever Florida's equivalent is.

The discipline notices for these sorts of organizations are always public. You can browse through the cases on their website [1]. Most of the enforcement actions against licensed professional engineers in Oregon seem to be for failing to complete their required professional development hours (e.g. [2]), though I did find an interesting case of a P.Eng. who was sanctioned for revealing client data [3].

[1]: https://www.oregon.gov/osbeels/rulesstatutes/Pages/Disciplin... [2] https://www.oregon.gov/osbeels/Documents/FinalOrders/2019111... [3] https://www.oregon.gov/osbeels/Documents/FinalOrders/2019071...

Yeah, looking at the lists of violations makes it pretty clear that the system is obsolete. These are the same people that claimed they could fine every single engineer working at Intel for not technically being certified.

I'm sure that if they actually started going after people, everyone's job title would magically shift to "Technology Developer" or something similar.

Countless.

Here's [1] an example of the Board pursuing a company and their Professional Enigneers for an org structure that facilitated signing off of engineering work without satisfying the 'direct supervision' clause of the act. Specifically, the guy signing it off was based about 1500 km away from the team doing the work, and the head of the team doing the work, who was in reality supervising it, was not qualified to do so as a Professional Engineer.

As a second example, the coroner's investigation into the Dreamworld fatalities from a few years ago came out just last week [2,3]. In it the coroner recommended that the Board of Professional Engineers investigate the engineering firm that signed off the certification for the ride, over potential gross failures of practicing with due diligence. This may result in the approving Professional Engineer having their license to practice revoked.

[1]: https://www.queenslandjudgments.com.au/case/id/76335

[2]: https://www.abc.net.au/news/2020-02-24/dreamworld-accident-i...

[3]: https://www.smh.com.au/national/queensland/dreamworld-inques...

Are there jobs, other than being a judge or politician, where you are not liable for errors?

Software engineer

Software engineers, like everybody else, are absolutely liable for errors. However, their work is invariably licensed using a contract that disclaims liability.

Regulated professionals, by contrast, are often limited by law from disclaiming liability or otherwise shielding themselves from liability (e.g. by using a corporation). This is especially true for highly regarded professions like lawyers and doctors. Professionals and licensed workers more generally are also usually required to be bonded or carry insurance so that they're not judgment proof.[1]

[1] The reason you would rarely see a software engineer employee (or any employee) sued for negligence is that it would be a waste of money--they wouldn't be able to to pay a judgment. You sue the employer not because the employee isn't liable, but because 1) it's easier to prove a case against an employer and 2) the employer is the only one able to afford a judgment. For example, if a UPS truck driver runs you over, you always sue UPS. You may or not may not sue the truck driver, but unless the truck driver is secretly rich it's UPS you're expecting to pay out.

To cue up the old saw: Weathermen.

"You can be wrong half the time and still have a job" har har har

Appeal to authority is generally not a healthy line of arguing. Only the argument itself should be considered.

A valid choice if you have the time to make yourself enough of an expert on any argument that gets presented to correctly understand and judge its contents. Most people do not have that time or even the patience to waste it when we have various ways to identify people that should know what they are talking about.

If you do not use the validity of the argument itself, then you will keep ending up in this very same situation.

Remember, the original system was caused by "experts" in the area, whose authority far exceeded the person claiming the system incorrect.

Using authority as a measure would have lead to continued failure. It is always an error to use authority as proof.

Is that mentioned somewhere? As far as I am aware red light cameras are often abused as money makers and cities have been known to reduce wait times in order to maximize profit over the safety of drivers. So the "experts" in question most likely wouldn't have been engineers, they would have been bean counters.

The Professional Engineer title in much of the United States is an additional level of credentialing and is more rigorous than just completing an undergraduate degree in an engineering field (though I guess some areas let you qualify with years of regular demonstrable experience).

The Engineer license came up particularly to address quality control in civil engineering works [1]. So it's actually important to protect that, lest a bunch of folks start deciding they're qualified to give opinions on complicated structures or earthworks.

1: https://en.wikipedia.org/wiki/Regulation_and_licensure_in_en...

Yes but the man didn't claim to be a registered PE, qualified and entitled to practice in Oregon, he simply said that he has an engineering degree and is employed as an engineer. There are many many people who are engineers but who are not registered Professional Engineers.

Your description of his description is easily confused with him saying he’s an Engineer.

That’s a licensed position when you’re providing services to the public.

I think we’re desensitized by the software industry’s tendency to use the term for developers and technicians. “Software engineer” is not the same thing as an Engineer in the civil or electrical trade, for example.

No, it's not confusing at all. The guy never offered engineering services to the public.

This isn't anywhere near the same thing as "Software Engineer"

We're discussing this because he went and spoke at a public forum, and gave his expertise as foundation for why his speech should be influential. That can indeed be confused, and his obtuseness with not modifying his statements after multiple requests by the licensing organization would indicate that he meant this to stand.

The public are not experts on all things. Someone ignorant to the stages of licensing are likely to hear "I studied engineering and work as an engineer" and think, "I can trust this person with topics related to engineering." This used to happen, to undesirable effect, which is why calling yourself "an engineer" especially in civil engineering fields requires a license.

My problem with that has always been ...

If they were actually going out and enforcing that law generally, then it would have been reasonable. Maybe it would be a stupid law, but they'd be even-handed about it.

But, there's like, a ton of people you can find in five minutes on LinkedIn who are using "engineer" in the same way. They could have easily cracked down on that.

Instead, they leave the law mostly unenforced, and then throw the book at you the moment you make the government look bad with a well-documented study of traffic lights. What a crock. That's a blatant abuse of power.

That’s extra frustrating because if you think about the dynamics of an intersection, the person going straight through or making a left has a much higher chance of collision than someone making a right. There’s a longer interval before the perpendicular traffic intersects the path of someone turning right, and usually all they have to do is take their foot off of the accelerator, if that.

Interesting. Here in Australia the red light cameras only trigger upon entering the intersection - if you’re already in the intersection then that’s fine. I wonder what the solution they’ll implement will be?

Surely the rules in the US would have to be the same; you can't enter the intersection on a red, but if you're already in there you can exit. Yellow is meant to be "only enter if you can't stop safely" not "if you enter now you'll get a fine if you don't speed up"

Is Mats arguing that the slow down needed for the right turn could result in you entering the intersection as the light turns red?

It varies by state. In California, you can get a ticket if the light is red when you enter the intersection. In Oregon, you can get a ticket if the light is red when you exit the intersection.

Then what are you supposed to do if the light turns red while you're still on the intersection? Linger on the intersection until it turns green again?

I realise the purpose of the lights is to keep the intersection clear for the flow of traffic that has the green light, but in real life traffic situations, things rarely work out that way and some common sense is always needed.

Obviously my wording was poor. If the light turns while you are still in the intersection, you've earned a ticket.

In both states, you are supposed to not enter the intersection unless you have both time and space to make it through. In Oregon, intersections tend to stay clear more, because people generally follow that rule more. In California, if the light is yellow and traffic is stopped on the other side, many people will enter anyway, knowing they can't go through, just to claim their "turn". And totally block traffic when the light changes. Not so much in Oregon.

https://web.archive.org/web/20200301140653/https://www.koin....

https://archive.md/Sy521

https://outline.com/TtHBpK

In Australia you only get a ticket if the rear axle passes over the line while the lights are red.

It is very common for people to bank up an intersection and join the perpendicular traffic. Also we have left-lane-to-turn-right intersections.

Realistically being in the middle of an intersection when a light is red endangers no-one. You’re already there, so people can see you and aren’t going to T-bone you. Even if stationary.

> Realistically being in the middle of an intersection when a light is red endangers no-one. You’re already there, so people can see you and aren’t going to T-bone you. Even if stationary.

At low speeds, maybe. At anything above 30 mph, or at night, or with otherwise reduced visibility, it's still very dangerous to be in the middle of an intersection when traffic is flowing the other way.

There's also no right-turn-on-red setup in Australia, which means that this situation is unlikely to happen in the first place.

We have left turn on red, remembering we drive on the left hand side. It is not universal though, it has to be signposted.

In practice, you only get "caught" inside the intersection when there is heavy traffic, so speeds shouldn't be particularly high.

Whoops, you’re right. That would explain why I got a resounding “no” when I searched for “right turn on red Australia”. My American ignorance at work again :)

In some U.S. jurisdictions (like mine) you can turn left on red, when turning from a one-way street onto another.

If you can’t see into an intersection, common sense dictates to slow down because you won’t be able to see either the lights on the opposite side or the person driving in front of you.

There doesn't seem to be much information as to the legal basis for Järlström’s efforts -- is there some Oregon or US requirement that red light cameras need to be "fair" in order to be legal?

(Also -- would this extend to California? I received a red light ticket at an odd double intersection here I'm convinced involved a no-win situation, trying to figure out options, traffic lawyers of the ticket clinic variety do not seem to be of particular help.)

In this case, it's not a question about the camera at all - the camera was just what got him interested. The question at hand was the timing of the lights themselves, which were timed upon a simplified model that is not applicable to intersections with turns.

Well this is convenient timing.

I just received a citation for making a right turn at a red light 0.3 seconds after the light turned red. I had already been preparing my arguments for why my actions produced the safest possible outcome, but these materials should prove exceptionally useful in our case.

If it’s safe to reveal so, what made it safer to turn than to stop in your case? Having a hard time picturing such a scenario.

If the period of the yellow is too short, you can't safely stop in time.

Driving 30MPH. The light turned yellow when we were 100 feet from the intersection. If you are traveling straight through the intersection, you can be up to 154 feet away when the light turns yellow and make the light before it turns red.

As we decelerated to make the turn, I realized we might not make the light and had to decide whether to take the turn at an aggressive speed or heavily brake to a stop. Suddenly stopping would pose a significant risk to the passengers in my vehicle and in the vehicle behind me. I observed that the intersection was clear and committed to the turn.

This is the exact 'dilemma zone' described by Mats. If you are still having a hard time picturing it, there are plenty of diagrams in the ITE journal article that explains his calculations.

Referenced journal article, which hopefully is available to all:

https://www.nxtbook.com/ygsreprints/ITE/ITE_March2020/index....

Previous discussion (2017):

https://news.ycombinator.com/item?id=14234223

Also https://news.ycombinator.com/item?id=14197512

Red light cameras are nothing other than a revenue machine. Be suspicious of anyone who advocates for them.

Enforcement of traffic rules shouldn't have to be a problem, but when fines are seen as revenue, that may create perverse incentives for the people who implement the rules. Yellow lights staying yellow for just a bit too short is one of them.

Wasn’t this guy fined for “practicing engineering without a license” or something similar?

Indeed he was, and it shouldn't really need to be a cautionary tale because that's the natural result of giving organizations like this such power.

It's why software is so much better: no one takes certification guys seriously.

Ah, yes. Software is so much better. https://xkcd.com/2030/

Killed fewer than any of the others.

Because we rarely put it in a position of power. When we do, stuff like the 737 Max (where badly-programmed software attempted to compensate for badly-designed hardware, and probably could've got away with it if it weren't badly-programmed), or the 1994 Scotland RAF Chinook crash (where the software controlling the engine was just rubbish, but it's unknown whether that actually caused the crash), or Therac-25 (where the software controlling the radiation therapy machine was rubbish) happens.

If you're counting war systems, the list is much longer. I don't.

All of these issues were caused by ordinary software development practices being applied in a context where the software was actually important. Real-world software development practices completely suck, and have to be completely thrown out of the window if everybody's going to get out alive.

The reason more people haven't died is that software isn't given positions of responsibility very often, and it's usually held to actual standards when it is.

And that's what good engineering is. When you know your product can't do the job, you don't put it in a position of responsibility. If you can't do it safely, don't do it. That's good engineering.

Physical world engineers could stand to learn a lot from software engineers about the hard parts of engineering. Sometimes it's about saying "No, we shouldn't."

Yes, but he won that first-amendment lawsuit in late 2018: https://www.documentcloud.org/documents/5671551-jarlstrom1.h...

Sorry but this somehow strikes me as a very swedish thing to do.

traffic cameras are put up to collect rent, not enhance safety :(

Would formal verification be helpful in this situation?

How? This is a question of policy, really.

Whether or not a driver would find themselves in a situation where they would get a ticket despite following the law. Basically where they would have to run a red light.

Archived version, since access is blocked for european readers: http://archive.is/Sy521

The original article is not available to me because "European Union visitors are important" to them, the archive version, miraculously now compatible with my dns service, features a Google captcha which is unsolvable due to being half off the page.

Where did it all go wrong?

There wasn't a singular moment when it went wrong. It's been a long, steady decline. Hackers have gradually been marginalized and replaced by bureaucrats.

I've been reading Hackers: Heroes of the Computer Revolution [1] by Steven Levy. It's fascinating, but also heartbreaking, to see how things have gone, given that the hackers of the 60's and 70's were fighting against it even back then.

[1] https://en.wikipedia.org/wiki/Hackers%3A_Heroes_of_the_Compu...

Well, to be fair, the transition between the hackers and the bureaucrats was greased by abusive MBAs.

EU data privacy legislation wouldn't have been seen as necessary if companies weren't exploiting their user base unethically. The fight to keep the Internet open to everyone should have been an easy one for the "hackers" to win... but the other side had help.

Yup, all those FAANGS were founded and are currently run by evil MBAs. In reality of course, even Cook started out with a STEM degree.

Indeed, it's sad to behold. That looks a good book mind, thanks.

Anyway, as this isn't Bureaucrat News, I think it would be good if we all started to clamp down on links to articles on the quasi-web.

"Your privacy is real important to us and we would like your consent to rape it and share its remains with our partners."

We know, don't we? In the past, people always knew there was a value exchange occurring when you got info. Now, everyone wants the info for free and complains when people don't want to give it to you. Well, that's the nature of systems with this kind of incentive flow.

Really, targeted ads were a low-cost way of providing value to content producers which enabled a low-cost way for them to provide to consumers.

Fixed cost barriers increase the cost of doing business with these predictable results. It's like creating a slope and then complaining when the ball you placed on it rolls down. That's just the nature of balls and slopes.

Either

"Our European visitors are important to us."

"This site is currently unavailable to visitors from the European Economic Area while we work to ensure your data is protected in accordance with applicable EU laws."

Reply on news.ycombinator.com