[potato3732842]

The problem is that even if they look back and fourth and know you're there the "go" condition (no incoming cross traffic) is the same for both parties so it's a ready made "two idiots trying to pass each other in the hallway" situation.

I think it speaks volumes that the discussion is anchored around whether cars look or not despite the fact that the underlying algorithm will produce conflicts even if they do.

[ajuc]

The algorithm (if followed) does not produce collisions. Pedestrians have the priority, and in many countries (inlcuding Poland where I live) cars have to stop even if nobody's on the pedestrian crossing yet - it's enough that pedestrians are approaching the crossing.

This has changed in the last 10 years in Poland, and there have been numerous angry debates. It was introduced anyway, and the safety improved.

It's only a problem if we let drivers get away with making it a problem. The inherent asymmetry in the driver-pedestrian relationship must be taken into account by the law and road design.

[potato3732842]

>The algorithm (if followed) does not produce collisions. Pedestrians have the priority,

Yes, in magical textbook land sure. In reality there are signaled crosswalks and most pedestrians abide by them so it's not clear if any given pedestrian wants to cross at that time and the pedestrian is also looking for traffic coming from the right if they're crossing against the signal (perfectly legal, but ill advised in the face of social norms) it's a recipe for confusion. Multiply by a nation of hundreds of millions and you get a lot of near misses and accidents.

>It's only a problem if we let drivers get away with making it a problem. The inherent asymmetry in the driver-pedestrian relationship must be taken into account by the law and road design.

I propose a 3 step solution to this "problem":

1)ignore anyone who talks like that from any side of the issue because they're probably gonna make it worse and not better and piss everyone off in the process and make the problem harder to solve.

2) Slap up "no right on red" signs and adjust signals accordingly

3) Measure results and address gaps.

[ajuc]

> In reality there are signaled crosswalks

On signaled crosswalks, it obviously only applies when the light is green for pedestrians. Somebody's near the crosswalk and they have green light = you stop. It doesn't matter if they want to cross. Simple as that.

> Multiply by a nation of hundreds of millions and you get a lot of near misses and accidents.

Nation size doesn't matter for this. Poland based this law on experience from Lithuania which is 20 times smaller than us. It worked for Lithuania and it worked for us. Why would it suddenly be worse for 350 million people if it worked for 2 million and 40 million?

> ignore anyone who talks like that from any side of the issue because they're probably gonna make it worse and not better and piss everyone off in the process and make the problem harder to solve

When they teach you to optimize polynomials at school, they tell you to look for zeroes of the derivative and check which one is the global maximum. But they also tell you to look at the edges of the domain, because the highest peak might be outside the domain altogether.

I'd argue that the US is so car-centric that any effective solution will be outside of the perceived "practical" domain.