dkbrk · HackerTrans

一昨日議論

Yes, it matters. I found your comment hard to read and had to re-read it a couple of times to parse it. If you want to use the hyphen because it's more easily typed I suggest using the double hyphen "--", which perhaps doesn't have the correct semantics on a per-character basis, but at least looks mostly right and will be clear in meaning to a human reader.

dkbrk

23 日前議論

Could you actually explain what you disagree with? In my opinion, everything in the comment you're replying to is obviously correct.

If you're going somewhere where there is a chance you might get lost, injured or trapped by weather, and need rescue, you should already be bringing something like a Garmin inReach. That's a highly ruggedized device with a battery that lasts for over a week without recharge, is small enough to keep in a pocket, provides two-way messaging and weather reports, can track your position at regular intervals so your family can see where you are, and can, without any setup and even when you're seriously injured, be used to directly send out an SOS with automatic reporting of your position and two-way voice communication.

As excellent as Starlink is, it is nowhere near a substitute for those capabilities. And the inReach has existed for longer than Starlink, ergo Starlink doesn't change the risk profile. The only real argument that Starlink changes the risk profile is if you're comparing Starlink vs nothing, or Starlink + PLB vs just PLB. And sure, in those cases Starlink is a significant improvement, but it's still inferior to something like an inReach.

The second part of the argument is that having better connectivity is no substitute for fundamentals, which is overwhelmingly, obviously correct. Yes, bring all the connectivity you want, the more the better if you're willing to carry it. But your plan shouldn't be built around the assumption that you can be rescued if things go wrong. If you get complacent due to having better connectivity it's entirely possible for it to worsen, rather than improve, your risk profile.

dkbrk

先月議論

You've got that completely backwards. Correctly applying Bayes' theorem, if an anomaly is observed you incorporate the prior into the calculation of the posterior probability. You don't just give up and say "the prior is miniscule so the likelihood is useless".

And then, even that's not enough. Decision theory needs to be applied to decide what action to take. That means taking into account the expected QALYs, cost and inconvenience across the distribution of possible outcomes. There's a whole spectrum of possible decisions, from immediately performing surgery, performing an invasive test like a biopsy, performing other less invasive tests, scheduling a follow-up non-invasive test at a later date, or just following a regular schedule of non-invasive tests and looking for any evolution along a longer time period.

The real problem is the binary thinking of either "we think you have X" and therefore tests must be performed or "we think you don't have X" and therefore tests shouldn't be performed. If medical organizations adopted empirically grounded decision frameworks, by applying them consistently doctors would be able to see something anomalous, assess that the risk isn't high enough to warrant immediate investigation, and be protected from a lawsuit in the unlikely case it was, in fact, something. And then we could do away with this "if we look we might find something" nonsense, which is pure fallacy.

dkbrk

6 か月前議論

You can look at the Wikipedia page on railway defect dectectors [0].

Under "rail break monitors" it mentions both electrical continuity and time-domain reflectometry can be used, and are most frequently used on high-speed tracks.

In addition, there are vast array of other detectors using acoustic sensors, strain gauges, accelerometers, cameras in the visible and infrared spectrum or laser measurement, that potentially could have detected an anomaly (i.e. damage to the wheels of other trains before the incident).

[0]: https://en.wikipedia.org/w/index.php?title=Defect_detector

dkbrk

7 か月前議論

That's not what "hot" means in this context. "Hot" means "highly radioactive", i.e. high number of decay events per second, high concentration of short half-life isotopes, high power/volume resulting from radioactive decay.

Nuclear reactors do not work off radioactive decay. U-235, for example has a half life of 704 million years. Radioisotope thermal electric generators [0] by contrast do run off radioactive decay, an isotopes used for that application have short half-lives, such as Pu-238 with 87.7 years.

Commercial nuclear reactors use unenriched or minimally enriched fuel. This means that, within a fairly short period of time, the percentage of fissile material in the fuel drops to the point where continuing to use it is no longer economical. At that point the fuel is a mixture of extremely hot fission products, transuranics, unreacted fuel, and non-fissile (but fertile) isotopes such as U-238.

It's not practical to use the decay energy from the fission products for power. What would make much more sense would be to remove the fission products and recycle the fuel that remains into new fuel (for a reactor that's designed to use it). This would be a much more efficient use of mined nuclear fuel (allowing nuclear power to be used for thousands of years), it would vastly reduce the volume of nuclear waste, and it would mean nuclear waste would only be hazardous for decades to centuries.

The US was on the path to this with the Integral Fast Reactor and Pyroprocessing [1] developed by the Argonne National Laboratory. This was killed [2] in 1994 by the Clinton administration. Not for any technical reason, but because it was a "threat to nuclear non-proliferation". How that makes sense when, to the best of my knowledge the process developed by Argonne couldn't be used to produce weapons-grade material, and even if it could the US already had nuclear weapons so it wouldn't be proliferating it to a non-nuclear country, I don't know. But, apparently, since some other forms of nuclear waste reprocessing can be used to generate weapons-grade material (by extracting Pu-239), it was a bad symbol so it had to go.

[0]: https://en.wikipedia.org/wiki/Radioisotope_thermoelectric_ge...

[1]: https://en.wikipedia.org/w/index.php?title=Integral_fast_rea...

[2]: https://en.wikipedia.org/w/index.php?title=Integral_fast_rea...

dkbrk

7 か月前議論

https://slatestarcodex.com/2014/09/30/i-can-tolerate-anythin...

dkbrk

8 か月前議論

Your question is rather ambiguous. Do you mean using chemistry to develop new techniques or combine unusual ingredients to create food that has novel flavors or textures? That would fall under Molecular Gastronomy, which has been highly influential within fine dining in the last few decades.

Do you mean processing ingredients with the goal to take cheap ingredients and make a product as hyper-palatable as possible? That would generally be called "ultra-processed food"; you're not going to find a Doritos chip in nature.

Do you mean developing completely completely new flavors via chemical synthesis? I don't think there's much possibility there. Our senses have evolved to detect compounds found in nature, so it's unlikely a synthetic compound can produce a flavor completely unlike anything found in nature.

Also, I think you're overestimating jelly. Gelatine is just a breakdown product of collagen. Boil animal connective tissue, purify the gelatine, add sugar and flavoring and set it into a gel. It's really only a few of techniques removed from nature. If you want to say it's not found in nature, then fair enough, but neither is a medium-rare steak.

dkbrk

8 か月前議論

Have you seen the presentation from GDC 2017 on the architecture of Overwatch [0]? If you watch the video in detail -- stepping through frame-by-frame at some points -- it provides a nearly complete schematic of the game's architecture. That's probably why the video has since been made unlisted.

[0]: https://www.youtube.com/watch?v=W3aieHjyNvw

dkbrk

9 か月前議論

I'm not aware of anything quite like that, but most submarines have something like a Rescue Buoy [0], Submarine Emergency Position-Indicating Radio Beacon (SEPIRB) or Submarine Emergency Communications Transmitter (SECT). I think those might differ based on whether they're attached by a cable and allow communicating to the submarine, or just broadcast a distress signal with the position. In any case, they're designed to be automatically deployed in the event of an emergency or catastrophic event, and based on this Quora answer [1] they're attached by an independent mechanism with a timer which has to be regularly reset to stop it deploying. I think it might be a clockwork mechanism, with an electronic alarm when it's about to go off to remind the crew to wind it.

[0]: https://en.wikipedia.org/wiki/Rescue_buoy_(submarine)

[1]: https://www.quora.com/Don%E2%80%99t-submarines-have-communic...

dkbrk

昨年議論

Thanks for the link, but I'm not sure what the point of the 50 minute video is. Here's [0] the pdf of the report. It's really not that long.

[0]: https://drive.google.com/file/d/1L7RXrneHlzfjrewMFIeeyc-nel3...

dkbrk

昨年議論

It's a bit of a long read, but I think the best introduction is still this [0] and the comments were here [1]. Yes, it's presented in the context of rust and gamedev, but ECS isn't actually specific to a particular programming language or problem domain.

[0]: https://kyren.github.io/2018/09/14/rustconf-talk.html

[1]: https://news.ycombinator.com/item?id=17994464

dkbrk

昨年議論

> I'm not sure about language design or system architecture but this is almost universally not true for any mathematical or algorithmic pursuit.

I don't agree. While starting with the simplest case and expanding out is a valid problem-solving technique, it is also often the case in mathematics that we approach a problem by solving a more general problem and getting our solution as a special case. It's a bit paradoxical, but a problem that be completely intractable if attacked directly can be trivial if approached with a sufficiently powerful abstraction. And our problem-solving abilities grow with our toolbox of ever more powerful and general abstractions.

Also, it's a general principle in engineering that the initial design decisions, the underlying assumptions underlying everything, is in itself the least expensive part of the process but have an outsized influence on the entire rest of the project. The civil engineer who halfway through the construction of his bridge discovers there is a flaw in his design is having a very bad day (and likely year). With software things are more flexible, so we can build our solution incrementally from a simpler case and swap bits out as our understanding of the problem changes; but even there, if we discover there is something wrong with our fundamental architectural decisions, with how we model the problem domain, we can't fix it just by rewriting some modules. That's something that can only be fixed by a complete rewrite, possibly even in a different language.

So while I don't agree with your absolute statement in general, I think it is especially wrong given the context of language design and system architecture. Those are precisely the kind of areas where it's really important that you consider all the possible things you might want to do, and make sure you're not making some false assumption that will massively screw you over at some later date.

dkbrk

昨年議論

I actually think that it does a disservice to not go to Nazi allegory, because if I don't use Nazi allegory when referring to Oracle there is some critical understanding that I have left on the table; there is an element of the story that you can't possibly understand.

In fact, as I have said before and I emphatically believe, if you had to explain the Nazis to somebody who had never heard of WWII but was an Oracle customer, there's a very good chance that you actually explain the Nazis in Oracle allegory.

So, it's like: "Really, wow, a whole country?"; "Yes, Larry Ellison has an entire country"; "Oh my god, the humanity! The License Audits!"; "Yeah, you should talk to Poland about it, it was bad. Bad, it was a blitzkrieg license audit."

https://www.youtube.com/watch?v=79fvDDPaIoY&t=1459s

dkbrk

昨年議論

The discussion near the end about how leadership taking responsibility can beneficially relieve accountability reminded me of the story of the Naval Tactical Data System (NTDS) [0].

[1]:

> When NTDS was eventually acclaimed not only a success, but also one of the most successful projects in the Navy; it amazed people. Especially because it had stayed within budget and schedule. A number of studies were commissioned to analyze the NTDS project to find why it had been so successful in spite of the odds against it. Sometimes it seems there was as much money spent on studying NTDS than was spent on NTDS development.

[2]:

> ...the Office of the Chief of Naval Operations authorized development of the Naval tactical Data System in April 1956, and assigned the Bureau of Ships as lead developing agency. The Bureau, in turn, assigned Commander Irvin McNally as NTDS project “coordinator” with Cdr. Edward Svendsen as his assistant. Over a period of two years the coordinating office would evolve to one of the Navy’s first true project offices having complete technical, management, and funds control over all life cycle aspects of the Naval Tactical Data System including research and development, production procurement, shipboard installation, lifetime maintenance and system improvement.

[1]:

The Freedom to Fail: McNally and Svendsen had an agreement with their seniors in the Bureau of Ships and in OPNAV that, if they wanted them to do in five years what normally took 14, they would have to forego the time consuming rounds of formal project reviews and just let them keep on working. This was reasonable because the two commanders were the ones who had defined the the new system and they knew better than any senior reviewing official whether they were on the right track or not. It was agreed, when the project officers needed help, they would ask for it, otherwise the seniors would stand clear and settle for informal progress briefings.

The key take-away is that the NTDS was set up as a siloed project office with Commanders McNally and Svendsen having responsibility for the ultimate success of the project, but other than that being completely unaccountable. There were many other things the NTDS project did well, but I believe that fundamental aspect of its organization was the critical necessary condition for its success. Lack of accountability can be bad, in other circumstances it can be useful, but diffusion of responsibility is always the enemy.

How many trillions of dollars are wasted on projects that go overbudget, get delayed and/or ultimately fail, and to what extent could that pernicious trend be remedied if such projects were led from inception to completion by one or two people with responsibility for its ultimate success who shield the project from accountability?

[0]: https://ethw.org/First-Hand:No_Damned_Computer_is_Going_to_T...

[1]: https://ethw.org/First-Hand:Legacy_of_NTDS_-_Chapter_9_of_th...

[2]: https://ethw.org/First-Hand:Building_the_U.S._Navy%27s_First...

dkbrk

昨年議論

Did you not read the paper? That's in there.

dkbrk

昨年議論

CNC Kitchen put out a great video on the practical use of silica gel. I especially found his exploration of different methods of drying to be of interest.

https://www.youtube.com/watch?v=6tHInlFfMcM

dkbrk

昨年議論

There's an excellent publicly released video made about an older test. It has some of the clearest footage of the actual reentries and impacts, and is also worth watching for the sheer 90s vibe of the production.

https://www.youtube.com/watch?v=PDL_pIPScSI

dkbrk

2 年前議論

This is a bit late, but I don't see any other answers that provide what I think is the key insight.

The accounting equation is: Assets = Equity + Liabilities.

For a transaction to be valid it needs to keep that equation in balance. Let's say we have two asset accounts A1, A2 and two Liability accounts L1, L2.

A1 + A2 = Equity + L1 + L2

And any of these sorts of transactions would keep it balanced:

(A1 + X) + (A2 - X) = Equity + L1 + L2 [0]

(A1 + X) + A2 = Equity + (L1 + X) + L2 [1]

(A1 - X) + A2 = Equity + (L1 - X) + L2 [2]

A1 + A2 = Equity + (L1 + X) + (L2 - X) [3]

Now, here is the key insight: "Debit" and "Credit" are defined so that a valid transaction consists of the pairing of a debit and credit regardless of whether the halves of the transaction are on the same side of the equation or not. It does this by having them change sign when moved to the other side.

More concretely, debit is positive for assets, credit is positive for liabilities. And then the four transaction examples above are:

[0]: debit X to A1; credit X to A2

[1]: debit X to A1; credit X to L1

[2]: credit X to A1; debit X to L1

[3]: credit X to L1; debit X to L2

You can debit and credit to any arbitrary accounts, and so long as the convention is followed and debits and credits are equal, the accounting equation will remain balanced.

Another way of looking like this is with parity. A transaction consists of an even parity part "debit" and an odd parity part "credit". Moving to the other side of the equation is an odd parity operation and so a credit on the RHS has double odd parity, which means it adds to those accounts (and debit, with odd parity, subtracts).

dkbrk

2 年前議論

That's not what "indiscriminate" means.

Indiscriminate attacks are those [0]:

(a) which are not directed at a specific military objective;

(b) which employ a method or means of combat which cannot be directed at a specific military objective; or

(c) which employ a method or means of combat the effects of which cannot be limited as required by international humanitarian law;

and consequently, in each such case, are of a nature to strike military objectives and civilians or civilian objects without distinction.

The fact that the pagers were obtained by Hezbollah to be used for their communications, and consequently could be expected to be exclusively in the possession of combatants means the attack was not indiscriminate.

Causing collateral damage does not make an attack indiscriminate. The standard for permissible collateral damage is that an attack must not cause loss of civilian life, injury to civilians, damage to civilian property, etc. that is excessive in relation to the anticipated concrete and direct military advantage [1].

The fact that it was so specifically targeted, combined with the small size of the explosive charges means collateral damage could be expected to be minor. And the evidence so far suggests that to have been accurate. The death of a single child is tragic, but negligible in comparison to the military advantage gained by thousands of combatants dead or wounded.

[0]: https://ihl-databases.icrc.org/en/customary-ihl/v1/rule12

[1]: https://ihl-databases.icrc.org/en/customary-ihl/v1/rule14

dkbrk

2 年前議論

You use radiation for cooling in space. That obeys the Stefan-Boltzmann law with power scaling with T^4 (so you want the radiators as hot as possible). You can use passive elements, like heat-pipes to move heat to the radiators, but active elements like pumps for forced convection could make sense; and most importantly heat pumps are an active element that can boost the temperature of the radiators vs the thing you're keeping cool (thereby increasing the heat rejection capacity of a given radiator size).

コメント