Comment by krisoft | Hacker News Reader

Comment by krisoft | Hacker News Reader

Parent Full thread

> There was a rule in the Port of Los Angeles saying you could only stack shipping containers two containers high.

This is incorrect. There was a zoning rule which affected truck yards in Long Beach and Los Angeles. Truck yards. Not the port itself.

As stated in the linked tweets actually.

But if you don't believe that you can just google an image of the Port of Los Angeles from let's say 2019 and count how high the container piles go. Here is a randomly selected image from 2019 where 5 high piles can be clearly counted: https://www.joc.com/sites/default/files/field_feature_image/...

Accuracy is important. I'm not an expert on logistics, or zoning laws. But how could I trust the article's author when they clearly unable to parse their own sources?

> Normally one would settle this by changing prices, but for various reasons we won’t get into price mechanisms aren’t working properly to fix supply shortages.

It's nice that the article is not going into that. Instead it hammers on that politicians regulate where and how many containers can you plop down. That is not the real issue.

If you are moving containers into an area, and you are not moving an equal amount out then you are going to run out of space to store the containers. It is that simple. You can tweak rules to make a bit more space, for example by stacking them higher in the truck yards. But the real question is: why are the people who own these containers incentivised to move them back to where they want them to be filled? If you solve that the problem solves itself. If you can't solve that piles of containers will fill up what little more space you won by tweaking. So the very point the article decides to "not go into" is the only one worth going into.

I agree with everything you said.

Anecdote: I was driving into San Pedro in 2019, and I didn't have a smart phone at the time (so no map/gps). I took the wrong exit off of the 710 and ended up on Terminal Island. That was the most visually overwhelming place I have ever been... the scale of the ships, the height of the stacked containers (more than 2), the abundance of trains... the cranes... visually, overwhelming. And then there was all the road work, construction, detours, one-ways down wrong-way streets.... I was a hell of a morning as I tried to get to my presentation....

> is the only one worth going into

Sure, when your critical system goes down an RCA is hugely important and ultimately you have to apply a fix that addresses the core issue to avoid it happening again in the future.

But, at the time that the system is actually down it seems like the most important first step (once you understand the problem) is to get the system running again ASAP. This can give you the runway to fix the actual problem.

It sounded like the capacity to physically move things around is being blocked because trucks are being used for storage.

In that particular situation a temporary buffer that allows the flow to become unblocked is necessary.

The computer won't operate if you are unable to move data off the internal registers because there's nowhere to go. Including operations to delete the data in long term storage that is preventing the internal registers from being cleared.

For us software developers, it seems like the shipping industry has the container equivalent of a memory leak (a "container leak" if you will). Then the stacking rule change is the equivalent of simply adding more memory to the system, it doesn't fix the problem, but it buys you some more time of normal system operation before the next out-of-memory crash. Hopefully they use this time they bought to work on an actual solution to the original leak.

If you want to phrase it in programming terms, there's a pretty obvious comparison to semaphores and deadlocks. We've deadlocked. As a workaround we've increased a bunch of our semaphore limits. But we also had those limits for a reason (various balances of safety, efficiency, and available amount of other ancillary resources). Maybe now we'll run out of some other resource, or hit contention somewhere else - we don't really know, the system is big and hasn't been operated in this state before. And on top of that, we didn't solve the fundamental deadlock - if the underlying conditions persist, there's hard limits to how often we can do this before we deadlock permanently.

No, it isn't a leak as the containers are empty but not garbage. It is very common to allocate extra memory that isn't used - in garbage collected languages you often need to fight the garbage collector for maximum performance in specific ways - which means you will have empty objects just waiting to be filled. You will get around to them eventually, but for now they are just taking up memory. If you allocate more of these objects than you have physical ram you will start swapping - but there is no leak, you will eventually either use them, or destroy them.

If you don't worth in latency senstive applications you might not have encountered a situation where you need to apply the above tricks.

Lol!

Continuing this analogy, what has happened is that the swap space for containers has filled, and it now has a form of compression applied to it, so that five containers can now be stored in the space where two could be beforehand.

(Edit: let's hope the swap space doesn't become encrypted.)

I wonder when the out-of-memory-killer process will start up? What would it look like--just not shipping anything to the US for a few months?

Counterpoint, sometimes it's better to let the system burn, or else the root cause will never be addressed. Treating the symptoms can take the pressure off solving the root cause.

Correct! If we simply let the person die of cancer, we can properly investigate the tumor when they are dead.

I should have given more context. In cases where incentives are deeply, structurally misaligned, and it will take heroic effort and significant luck to yield an order of magnitude improvement over the status quo, we should consider "letting it burn" as an option, and recognize the total cost of treating the symptoms. The global logistics quagmire may be a candidate for nuclear-ish options. Agree with you on the cancer patient scenario.

Let the global supply chain burn? By Jove, let’s have a great depression!

Let the containers on the streets piss people off to build pressure to align incentives, rather than prolonging the problem with a temporary stacking improvement. This is just a tool in our toolbox that we should not ignore, I'm not saying it's the right tool. But there is a cost of papering over the root cause, that's not free.

BTW I don't live in LA/Long Beach. I recognize that LA doesn't deserve the quality of life degradation, that's an externality. We have tools to resolve externalities. I could imagine living in an affected neighborhood in LA and being super grateful for the container stacking "quick fix".

But the temporary rule change is a quality of life improvement. Less trucks idling on the streets, less ships idling offshore means better air quality. Less ships anchored means less chance of another oil pipeline spill. More cargo hauled means more prosperity. A few months of eyesore are an acceptable cost.

The thing is, there is no fix to the root cause. You can either have cheaper products with just-in-time supply chains, or you can pay more for storage. The trade-off will always be efficiency or redundancy, and most industries have already chosen the level of risk they can accept. Real world systems have tipping points and bottlenecks, and it’s okay to use government to push them back into steady state.

Why not? It worked last time.

The problem is that often the consequences of letting it burn are most felt by innocent bystanders, rather than the people who are meant to be "taught a lesson".

Assuming no other conditions change, how long will it take them to use up the extra storage space?

Indeed. Given the "balance" of trade, surely most of the empty containers need to go back on a ship so China can fill them up again? The problem is not the size of the buffer but the fact that we aren't emptying the buffer.

Yea, that stood out to me too the stacking limit was not at the port..

Ok, a fair point. But in the end is the same thing: stupid public worker bureaucrats exerting their petty, ignorant power like the Gods they think they are.

Unrelated to the trailer issue, but everyday we should thank our lucky stars that there's no such thing as "stupid private sector bureaucrats exerting their petty, ignorant power like the Gods they think they are."

You appear to be unfamiliar with the phenomenon of oil companies. Or power utilities.

PGE is considered directly responsible for >$40B in fire damage in California.

<I>PGE is considered directly responsible for >$40B in fire damage in California.</I> <p>That (officially sanctioned) conclusion was in complete disregard of the state and federal gov’t contribution. But I would agree with you that it’s an example of similar phenomena. Probably not for the same reasons.

IMHO regulation, the phenomena of regulatory capture, and the implementation of regulatory compliance bureaucracy within PG&E resulted in private sector bureaucrats drunk on power imposing government-supported misfeasance with unintended consequence...With a organization like PG&E it’s difficult to discern the boundary between the state and the corporation. Modern government enthusiasts dream of ways to impose more perfect order using the power of government, not realizing the emergent imperfection is a consequence of such aggrandizement.

Regulatory capture operates to eliminate the power of government to enforce responsible behavior in regulated industries.

Pretending otherwise is libertarian fantasy absolutely opposite to the objective facts.

Regulatory capture blurs the line between government and industry. There’s no fantasy about it. The power is suborned, not eliminated. However I was pointing out the phenomenon as a precursor to enabling petty corporate bureaucracy empowered with government authority.

With respect to your comment, are we supposing regulatory capture is never a responsible corporate behavior? I’m not sure how we must conclude that responsible corporate behavior necessarily follows from and is solely dependent upon the exercise of regulatory power.

OK, let's try again.

It is clearly possible, in principle, for a corporation to behave responsibly if its officers really want to and its board permits it. A corporation whose management wants it to do something responsible doesn't need a regulator to tell it that must. They can just do it.

What else, then, would lead it to choose to capture a regulator? To force its competitors to behave responsibly too? To force its future self to behave responsibly? Can you identify any single instance of either ever occurring?

Whatever may be possible, what we have seen over and over again is, instead, corporations capturing regulators and then hamstringing every effort to enforce any sort of responsible behavior.

Automobile manufacturers fought tooth and nail against requirements that they provide seatbelts, and then airbags, and then pollution controls, and then crash safety. Tobacco companies fought tooth and nail against labeling requirements, and restrictions on sales and advertising to children. Boeing management lately had the FAA approve their deathtrap 737-Max, at ultimately ruinous expense to their own stockholders and to airlines suckered into buying them.

The least harmful examples I know of have been to raise barriers to entry for their industry by imposing expenses that they, but not new entrants or smaller competitors, could afford, in the form of requirements on reporting, or fabrication materials, or quality standards, or occasionally even restrictions on effluents.

We see in many states a Dairy Council that has got itself delegated authority to assess their own taxes, spent then on billboards promoting dairy products, or buying up and destroying "excess" production, invariably favoring the biggest dairies and making smaller ones less competitive. Medical, dental, legal, hairdressing, and other "associations" are allowed to maintain licensing regimes to limit competition that, sometimes, act to establish a minimum required level of competency or education, but more reliably guarantee captive income for schools and exam boards.

You certainly have made some good arguments against regulation, suggesting a near certainty of regulatory capture screwing up whatever good intentions the original premise of regulation promised.

Was that your intention? I can’t dismiss your arguments out of hand as purely fantastic.

OnTheOtherHand, viewing all regulation and regulatory process as necessarily corrupt from birth doesn’t prevent some good coming of some regulation. I don’t think you’ll find corporations participating for the sake of highlighting their own immoral or irresponsible practices targeted by said regulation…maybe something along the lines of “thank god we helped develop this regulation that will allow us to stop doing these negative things forced on us by evil competitors/fraudulent bad actors/whatever scapegoat”. In spin world, everyone’s heroic in deed and motive…

Given your arguments, you might support this thesis: the more regulation and regulatory proceedings, the more likely some corruption may be concealed within it.

Thus a corporation seeking limits to regulation is acting responsibly to limit potential corruption, regardless of motive. It also follows that anyone seeking expansion of regulation is inviting corruption, regardless of motive (aka “the road to hell is paved with good intentions”, HaHaOnlySerious).

Obviously overtly corrupt and succinct regulatory proceedings are outside the scope of this thesis. Not that I’m aware of any succinct regulatory proceedings, heheh.

InMyHumbleOpinion the libertarian perspective would view using the power of government with good intentions as fraught with unintended negative consequences. Therefore we should use a minimum of government power exercised with maximal certainty of appropriateness. If we are not unanimously and honestly certain then government should refrain from the exercise of power.

I feel I might have left out something about how being cynical about regulation doesn’t make one an irresponsible actor but then this isn’t a retelling of “Anarchy, State, and Utopia”, “The Wealth of Nations”, “Atlas Shrugged”, or maybe “The Mystery of Capital”…

And, off we go to fantasyland. Again.

The rest of us have no reason to follow you on your flight into fantasyland. Have fun there.

You had the opportunity to take the high road...I mean, is that all you got? Seriously?

Idunno, to me it appeared more like perhaps you were unfamiliar with the phenomenon of sarcasm.

only by people who don’t understand the science of fires.

Fires need three things: fuel, oxygen, heat, and two are effected by humans.

Heat: PGE, Lightning, Vehicle fires etc. all statistically provide opportunities for heat.

Fuel: There is an order of magnitude more fuel than a century ago.

We went decades without PGE systems routinely causing fires. Conditions now are somewhat different, but not that different. The important difference is PGE's maintenance policies. They abandoned that responsibility in order to deliver a dividend. They were ordered to resume, but there is a huge backlog of work.

Reply on news.ycombinator.com