HNHacker News
TopNewBestAskShowJobs

cturner

4,781 karma · joined August 19, 2007

submissionscomments
cturner··on Python Asyncio
Different trade-offs. It is much easier to pursue cache-efficiency in a cooperative multitasking setting than with preempt threads.
cturner··on We will not pursue the potential acquisition of FTX
The toy model is the idea that a bank gets deposit from one customer and loans it to another. This is not how it works. If you go to a bank for a loan, they do a risk assessment of you, and then make some marks on a ledger. As a result of this, new money appears in your account.

But - what stops banks from creating infinite money? There is a byzantine system of rules laid out by the Basel Accords and national regulators within that. That system limits the activities a bank can conduct, lays out rules for the way it must go about business it can conduct, and sets capital requirements.

There is nothing close to a full reserve and the system is not defined in those terms. Rather, there are ratios. If you have so many AAA rated treasuries posted with your central bank, then you can grow your balance sheet by X. But for B rated investments, you can only grow your balance sheet by Y. When a commercial bank issues a loan, it literally creates money.

cturner··on The Essential Philip K. Dick
I have read a fair bit of pkd and found Valis an unsatisfying struggle. Radio Free Albemuth is an earlier attempt at similar ideas and is accessible.
cturner··on Can I use a system call?
Has any OS offered introspection so that you could determine such things at runtime?
cturner··on Acorn: A simple application deployment framework for Kubernetes
Interface Script http://songseed.org/post/20200607.aa.interface_script.html (n-dimension csv)
cturner··on Airflow's Problem
When we engage in complex work, it is important to keep our options open so we can direct our best effort to the hardest problems.

It is rarely clear what the hard problems will be when new to a domain. Only as scale kicks in.

We are constantly pitched frameworks that sell themselves as a good approach to a domain, but then obstruct engagement with the hardest problems when it matters. The developer becomes captive of the system that claimed it would steer them right.

This is particularly true of fields where the hard problems are integration problems which, by their nature, cannot be outsourced to frameworks.

cturner··on Rocky Linux 9.0
Driver compatibility with redhat. Niche hardware like stack bypass network cards, packet capture cards, fpga cards will tend to have drivers issued for redhat, but not always for anything else. By using rocky/alma, you improve your odds to dodge hassle adapting the driver packs to other linux distros.

Net booting. Distros vary in how awkward it is to get stripped setups set up for net booting. My experience is that the redhat world distros are easier for this than the debian world.

cturner··on FTC action against Harley-Davidson and Westinghouse for limiting right to repair
The concept of property evolved independently in every civilisation that grew beyond a small tribe and predates the nation-state.

It exists to deal with the dynamics of scarce goods. If person a is using your wheelbarrow, then person b cannot use it as well. You can't raise cows on a piece of land, but also plant a crop of wheat there. This definition goes back thousands of years and is respected across cultures.

There is another practice that was respected across cultures for thousands of years: the free sharing and open remixing of songs, stories and ideas.

IP is a weaselly project to carry the advantages of property rights to non-scarce goods, and to things which were previously fiercely regarded as public domain. It creates a set of rights which did not previously exist and which created new benefits for people who were already elites.

"For IP, others selling copies would restrict your ability to sell copies as they undercut you." This does not follow from the argument you are making. Selling something is not a use of it. Carting things around in a wheelbarrow or playing a compact disk - these are examples of use.

Perhaps you bring some cultural bias to this conversation. Having grown up in a setting with IP, and having a career in a field where some people gain great advantage from IP, you have come to believe that it is reasonable to regard ideas as property.

But IP is an aberration, even in our time. It is not law in all places, it is not enforced well in the places where it is, and in those places people who are morally conscious and otherwise law-abiding routinely violate it.

IP is impractical to enforce equally, is inconsistent with the principle of live-and-let-live, it raises the barrier-of-entry to a range of industries, it grants extra privileges to existing elites, it is easily gamed by bad-faith actors, it is disrespected by the general population, it has at best a hand-wavey economic justification that is backed by no real evidence. IP is complicated when good laws are simple, and IP muddies the water for a concept that is a genuine foundation of our civilisation - property. IP fails all tests for what is reasonable law.

If the government were to say that black were white, this would not make it so. So it is with property.

cturner··on Does the Tank Have a Future?
Not a direct answer, but an advantage of Israeli armour is that they only need to deploy locally. This means less penalty for armour than US systems which are designed to be carried in air transports for foreign deployment, or Russian units designed around train transport. The Israeli Namer IFV weighs more than twice as much as a Bradley IFV, and can carry more troops.
cturner··on Big money, nuclear subsidies, and systemic corruption (2021)
In this note you speculate optimistically in several places - "give it a few years", "I don't know what the long-term consequences of that will be", "I hope". Right now, we have what we have. The demonstrated paths to emissions reduction include nuclear and gas. We have a weird situation with electric vehicles in some regions where brown coal is being used to power electric vehicles, making for worse emissions at higher cost than oil - that's where speculation-based policy goes.
cturner··on Big money, nuclear subsidies, and systemic corruption (2021)
Pumped storage depends on geography, and it is not viable in most places. Hence, your claim that it is cheaper than nuclear is unreasonable.

> minus its free catastrophe insurance

Many more people have died from dam collapses in the last fifty years than nuclear accidents.

> Every kwh of solar and wind produced electricity right now > is a kwh not produced by gas and will be for some time > almost everywhere.

We don't have a general-purpose solution to storage. As a result there is waste overproduction of solar and wind at some times, and under-production at other times. Those under-production windows need to be covered by base-load forms of energy production such as gas and nuclear.

cturner··on How to walk underground from Liverpool Street to Farringdon
Accommodating the cheat, the journey could have been more elaborate: enter at Monument, walk underground to Bank, short hop on central line to Liverpool Street, then the rest of the journey.
cturner··on Finland’s Green Party endorses nuclear power
In general, solar is not in the same category of energy generation as nuclear because it can't carry the base load of the energy grid, because it is so inconsistent.

We do not currently have a general-purpose battery for solar and wind that could power the grid when they stop producing. (Some regions can through pumped storage hydro.)

Nuclear compares against base-load power sources: coal, gas, oil, hydro, geothermal.

cturner··on Finland’s Green Party endorses nuclear power
There is a comment against that video by user jem780 that claims to quote an Australian academic to counter the argument that there are inadequate uranium reserves.

When I was watching the video, that argument reminded me of the Peak Oil campaign from about twenty years ago. This claimed that the world had reached a point where the amount of oil available would be in decline going forward, and claiming that this justified major changes. At the time, this Peak Oil claim got a lot of sympathetic coverage in mainstream media. Oil production has been healthy in the years since, and technology changes have opened up new reserves.

cturner··on Shanghai video overwhelms Chinese censor, gets 400M views
Revolutions use angry young men as fuel. China's young men population is proportionally small. Further, due to the legacy of the four-grandparent-policy and Confucian respect-for-elders tradition, the young that exist have high family responsibilities. Perhaps there will be something new in China - something that comes from the same place as revolution but without the testosterone.
cturner··on DALL-E, the Metaverse, and Zero Marginal Content
If there is anyone here with access to it, please could you try using it to generate some tight pixel art of simple concepts and post the results? I am thinking - chair, coffee, boat, control panel, wedding ring, lamp, pit. I am curious whether it could be a fast way to design sprites. Imagine if it could be trained in the style of particular desktop systems.
cturner··on What is the Eurodollar System? (2020) [pdf]
This is not correct - "Consider what the logic of the Eurodollar system implies. Global financial markets and the global economy rely on the common standard of the USD for pricing, accounting, trading, and deal making. Imagine a world with a hundred different currencies – or even a dozen: it would be hugely problematic to manage, and would not allow anywhere near the level of integration we currently enjoy"

1. It is common to use USD because it is convenient, but it is not a standard. There is no structure that forces it.

2. The world does have dozens of liquid currencies already. Banks do offer rates on them. And banks do trade in them without the USD being in play.

3. It is simply not true that the current system of floating exchange rates would fail to function without the USD.

The author moves on to talking about it being a privilege to be a reserve currency. It is not a privilege to be the reserve currency - countries generally go out of their way to avoid it happening - the Swiss central bank headed it off within the last ten years. Patrick Boyle's recent video /The Weaponization Of The Dollar/ touches on this in the last few minutes.

I stopped reading here.

The concept of Eurodollars (US-denominated debt outside the reach of US foreign policy) has been effectively obsolete in the developed world for decades.

Any bank who has a balance sheet large enough to be involved in significant trading operations (e.g. operating as a top-level general clearing member at a clearing house, or acting as a prime broker for foreign exchange transfers) is also able to qualify for an account with the US central bank. This gives them access to cheap lending (free money).

When you take on this membership, you sign up to align with aspects of US foreign policy. And not just for stuff you do that is denominated in USD - it applies to all your business globally.

A recent example of this - Deutsche Bank got fined for money laundering done by Danske's Estonian Branch. Danske were operating on DB's balance sheet, which means that DB had failed to ensure that their balance sheet use aligned with US standards on money laundering, so DB got fined in the US even though they had no direct hand in the wrong-doing.

It is these arrangements that allow the US to be so effective in launching sanctions against other countries. Consider how Trump decided to go after Iran even though European countries wanted to stick by the Obama-era deal, and how thorough those sanctions have been.

Is it possible to have a top-tier western bank who choses not to have a relationship with the US central bank? Hypothetically, yes, but who would do such a thing? Such a bank would lose access to cheap lending in the world's largest economy and would lose economies of scale.

It is not the USD that is powerful, it is the US. The US is a vast economy with a strong rule of law tradition. Common use of USD is a side-effect of that.

cturner··on What is the Eurodollar System? (2020) [pdf]
A bank deposit is a debt, right?
cturner··on Managing People
"Beyond bootstrap, the company itself (revenue streams, assets, etc.) is the capital."

Thanks for this correction.

cturner··on Managing People
A difficulty with marxism is that it is thoroughly old fashioned about bucketing people into categories. In a feudal society you were born into a life where you were a serf, or a craftsman, or an aristocrat. Marxism massages the labels but essentially accepts this worldview. The primary handle on each person is a class label, assigned at birth. Talent is irrelevant.

One of the challenges of management after feudalism is making the best of the variety of individual contribution. When you are moving rocks around or ploughing, each individual is roughly equivalent. This was obsolete by the industrial era. A factor in Japan’s rise in auto was giving workers direct and individual (ie not union-filtered) feedback into the manufacturing process.

Now, the multipliers are higher still. There are regular situations where an aggressively small team has executed where several well funded teams of thousands failed.

My view is that capital is nearly irrelevant in mainstream technology now beyond achieving bootstrap and avoiding dumb renumeration errors that alienate talent. Rather, the key tensors are 1. how high you can get the ceiling on your core talent and 2. how small you can get that team, in order that this talent can be engaged with the big picture, in order that they can be identifying and focusing on the most important problems. They are related. The more capable your talent, the more it can participate in the big picture.

These dynamics do not resolve easily with the orgchart-centric, bigger-is-better assumptions of traditional worldviews.

Possible limit to the view - I have not had exposure to fields with vast ongoing capital cost, or complex cross-discipline stuff like silicon manufacture

cturner··on Reasons Kubernetes is so complex
The purpose of the shell is to allow users to interact with the grid. You can't do this in a conventional OS shell. Example: if you are using /bin/sh on linux-host-a and want to affect something running on linux-host-b, you can't. In a SSI shell, you can affect resources from a set of machines from a single shell.

You could offer something like shell scripting on that, but my system did not and there are better options. For example: have the resource manager expose a command that offers a DSL for making system-wide changes.

cturner··on Reasons Kubernetes is so complex
The paper presented declarative configuration as a contrast between the SSI approach and the k8s approach. I don't think it is.

It is useful to have both the fitted arrangements and operator interaction.

In addition to giving operators power to make changes in production, the command line is useful for development. Devs can develop in a one-host grid and interact with it in a repl-like way. Contrast this to traditional workflow where the deployment environment is quite different to the development environment. This encourages devs to work with the grain of the platform and discourages the "it works on my machine" dynamic.

cturner··on Reasons Kubernetes is so complex
Question about drivers. If you had DMA-oriented/kernel-bypass network cards or wanted to resource-manage GPUs, would a type 1 hypervisor be able to support this?
cturner··on Reasons Kubernetes is so complex
""" One could imagine a very imperative “cluster operating system,” like the above, which exposed primitives like “allocate 5 CPUs worth of compute” or “create a new virtual network,” which in turn backed onto configuration changes either in the system’s internal abstractions or into calls into the EC2 API (or other underlying cloud provider). """

There is a name for this: Single System Image. There is no need to write the operating system layer - Linux works well.

I built a SSI. Each grid can host several apps, 'gridapps'. As priorities change during the day, an operator can rebalance resources using a bash-like shell. Each gridapp has (1) core nodes, which need to keep running, and (2) scale nodes which can be started and stopped. Nodes can be configured so they are taskset-bound to an individual processor, or sharing a processor between several nodes. There are equivalent primitives for memory and persistent storage. Commercial, 1000 cores.

The author contrasts this with K8s by saying that K8s is declarative. This is not the important distinction. My system had declarative configuration, and did a fitting at startup.

The key difference is motive. The design above is the natural choice when you build a grid computer from first-principles.

K8s is designed to solve a different problem: scaling up the traditional model of programming. Here, a developer writes an algorithm on their workstation. The deployment circumstances and interactions with other systems are an afterthought. Often this developer does not deploy the code themselves, this is done by another team who don't understand the internals of the code.

cturner··on Ask HN: How to begin learning assembler for university course?
Some things that were high value to me,

1. Play through the first page of levels of TIS-100. This is a game you can buy on Steam. After spending a few hours working on problems in this, I felt a hard-to-describe mental shift that removed the intimidation of assembly.

2. Work through the first half of /Programming Boot Sector Games/ by Toledo.

3. Designing and building a simple instruction architecture, implementing it as a VM, writing an assembler for it.

The ChibiAkamus tutorials on youtube may be worth a look, also.

If you try for TIS-100, give yourself a decent chance to get into it. Print out the manual, calmly read through it with a biro, don't let yourself give up easily on the first few problems.

cturner··on What Do We Know About Time Pressure in Software Development?

  > Every deadline can be met if employees are "eager to
  > please" and will work unlimited overtime hours for free.
No amount of pressure or willpower can get software done if the team is badly aligned, working on the wrong problem or simply flailing. The message I take from the parent post is not about estimation theory, it's that the project manager created a delivery culture across a large team.
cturner··on Computers Built to Last
FeRAM may be viable for long-term rom, ram, and persistent storage. The usage contract should ensure it is recharged periodically. A plug-board could allow the user to manually populate the roms from paper if it lost charge.
cturner··on Fixing stutters in Papers Please on Linux
"implies non-blocking fds"

I don't think the open and close calls respect non-blocking on linux when operating on files (specifically files, not sockets - for sockets the return is always quick as far as I know). From man open(2), "I/O operations will (briefly) block when device activity is required, regardless of whether O_NONBLOCK is set". And my recollection is that they will non-briefly block if the problem is a hanging NFS mount.

"What are you going to do when they don't return? Accumulate dead threads and inconsistent shared application state?"

Good point. My practice is to use child processes. These can be killed, and so I do not run into this. But subprocesses ratchets up the amount of work to be done, because you then need to do async IPC. So it's now a lot of extra work. It's even worse for multiplatform stuff because now you are exposed to platform differences (e.g. select is suboptimal on Linux, but poll is not available on Windows).

As you say, using threads within same proc would lead to stale threads. In some contexts this would be tolerable but it is not nearly as simple+clean as I presented.

Thinking hard about addressing this on Linux brings me down, every time. I hope io_uring will make pure async practical within a single process. Even if it does, the multiplatform story will remain complex. I am not fond of your disregard for user data in the (tangential) discussion about disappearing filesystems, but you have won me over to embracing the main thread in this context.

cturner··on Fixing stutters in Papers Please on Linux
"You only need a mitigation if it ends up blocking long enough to be a problem in reasonable setups."

That criteria is established here - the OP is about an issue affecting paying end-users! Premature optimisation is not relevant - the software is failing.

It is unsafe to make fair-weather assumptions about customer systems.

Consider a common software failure: where the user is saving data to a SMB or NFS partition, and then there is a loss of connectivity to that file-server, and the developer has done whatever I/O they need on the main thread. This causes data loss.

You /should/ be able to assume (1) reliably fast return from core async-coordination syscalls (e.g. select, poll), and (2) that you will not suffer process or thread starvation caused by someone else. Respecting those constraints, it is good practice to isolate sync calls to a non-main thread in order to catch when they are not returning. This robustly covers both common scenarios (like the missing-filesystem) and obscure scenarios like the one in blog post.

cturner··on Buy a coal mine, drive a gas guzzler, and other uses of reverse logic
The purpose of economics is to explain complex interaction between systems. Analogy is a powerful communication tool. Such phrases allow us to communicate and reason about systems.

The concepts raise the standard of policy development. Consider how much easier it is to discuss a proposal thanks to 'The Tragedy of the Commons'. Those few words serve as a shorthand for a dynamic. Without it you might need to spend half an hour stepping people through the dynamic each time you wanted to refer to it, and you would lose people along the way.

← PreviousPage 5 of 34Next →