Thursday Throughputs: Captains in Ships Columbus Edition

Michael Siegel

Michael Siegel is an astronomer living in Pennsylvania. He blogs at his own site, and has written a novel.

Related Post Roulette

65 Responses

  1. Brandon Berg says:

    ThTh5: Okay, but what is actually going on in that table? There’s a footnote that says, “Interpretation of the case rates in vaccinated and unvaccinated population is particularly susceptible to changes in denominators
    and should be interpreted with extra caution,” but what exactly does that mean? What changes in denominators are there, and how do they explain these results?

    I’m pretty sure vaccination doesn’t increase risk of infection, but I can’t think of a good explanation for this.Report

    • Heavily vaccinated pops tend to be those most at risk (healthcare workers, etc.) and the vaccinated are much more likely to be tested — both as a routine part of their job and because their much more alert to this sort of thing. The studies in Table 1 test systematically.Report

  2. Brandon Berg says:

    ThTh8: There was a recent study in mice showing that injecting the mRNA vaccines directly into the bloodstream can cause myocarditis, and there’s speculation that accidentally hitting a vein could be what’s causing it in humans.

    If so, this could be prevented by aspiration (pulling out on the syringe plunger to see if blood comes out), which up until now has explicitly not been recommended.Report

  3. fillyjonk says:

    ThTh1: seems fitting to me now, that a guy who was in some ways “anti science of the time” (where people KNEW the earth was round but a westward route to Asia would take too long) insisted on his way, finally got it, managed to survive, and then some 450 years later was taught in primary school as a brave hero contrarian……I mean, maybe this is how we got into the mess we’re in now.Report

  4. Chip Daniels says:

    ThTh5, and ThTh9: Republican Jim Jordan has declared opposition to ALL vaccine mandates.

    Anti-vax hysteria is becoming the mainstream position for Republicans.Report

  5. Jaybird says:

    Okay. Classical Astrology.

    I have a theory about this. It’s a map/territory problem.

    Once upon a time, the majority of civilization all lived in more or less the same place more or less and more or less every year was the same as every other year… and how food impacts personality and the first few months of life outside of the womb either being swaddled from the cold or allowed to kick free in the warm sunlight would result in various personality traits. A mother who ate a great deal of ripe fruit during the first trimester and had a baby who spent the first month swaddled but after that was allowed to kick around would have these traits… and a mother who ate salted meat and preserved foods during the first trimester and went on to have a kid who had his earliest days wrapped up for warmth followed by months of being wrapped up for warmth would have an entirely different personality.

    “But what about”

    I’m not saying that there aren’t and wouldn’t be exceptions. Of course there would be. But there may have been enough there to generate broad-brush stereotypes.Report

    • fillyjonk in reply to Jaybird says:

      So it’s not “I’m needy and emotional because I’m a Pisces,” it’s “I’m needy and emotional because my mom was starved of sunlight for a bunch of the time she was carrying me.”

      It’s still woo, but it’s more PLAUSIBLE woo….Report

    • Kazzy in reply to Jaybird says:

      Ever since you first shared this, the premise has fascinated me.

      It isn’t entirely unlike Gladwell (I think it’s Gladwell?) look at why hockey players tend to have birthdays in certain months: it’s not because better players are born in those months but because having birthdays at certain times was advantageous in terms of the teams they were put on, the practice and development they got, etc.Report

      • Jaybird in reply to Kazzy says:

        Yeah. You’ve got tons of kids in your class. I’d guess that 1/12th of them have birthdays in March and another 1/12th of them have birthdays in October.

        Have you noticed any trends of differences between the two (otherwise pretty identical) groups of kids?Report

        • Mike Schilling in reply to Jaybird says:

          The kids born in December are less happy because they tend to get screwed on birthday presents.Report

        • Kazzy in reply to Jaybird says:

          Oh man, you’re asking the wrong guy. I’m terrible at remembering birthdays.

          I would also guess the broader trends that might have inspired astrology millenia ago have been muted now that we have things like climate control and year-round access to most foods and (in America, at least) most people don’t live lives that are dramatically dictated by the seasons.

          But since most schools have grades with 12-month age bands, we definitely see how age alone is a huge factor. It isn’t EVERYTHING because development is wonky BUT a kid who is 6 months older than another when you’re talking about kids who are 24-36 months old… well… that one kid has been alive for like, 20% longer than the other! How the hell can that not show up in some meaningful ways?Report

          • Jaybird in reply to Kazzy says:

            Oh, I want to say that pretty much everything except for Vitamin D has been wiped off the board since at, at the latest, the 70’s.

            And given the amount of time spent inside, that’s probably been neutralized as well.Report

  6. Oscar Gordon says:

    I saw this the other day. The tl;duc* is this; starch is an important food item. We get the vast bulk of our starch from plants. Plants take a while to grow, and the amount of usable starch that we get from any given plant is a fraction of the biomass. Growing such plants also requires a considerable amount of arable land and fresh water.

    To date, creating consumable starch without plants hasn’t been feasible.

    China says they have a new process, one that uses catalytic reactions to create food-grade starch, and also has the added benefit of consuming CO2. The starch process begins with the creation of methanol. I’m pretty sure I talked about this before, a few years back. So we know how to pull CO2 out of the atmosphere and lock it into methanol, but then what do we do with the methanol? This is the question the Chinese Academy is answering.

    Anyway, they still have to figure out how to build an industrial bio-reactor that would kick this stuff out by the ton, but it’s a pretty important breakthrough.Report

  7. veronica d says:

    [ThTh10] This is such a clever way to illustrate orbital resonance.

    On a related note, I recently discovered this Twitter account and it’s my new favorite thing: https://twitter.com/ThreeBodyBot

    As a diffeqs nerd, this was really fascinating. I wonder if, had 18-19th century people had computers, would they have given up the search for a closed-form and/or analytical solution earlier? Watching these unfold, it seems kind of obvious that no analytical function is going to describe these trajectories.Report

    • Oscar Gordon in reply to veronica d says:

      OK, that’s fun!

      We still haven’t given up on closed form solutions, I mean, the Millennium Prize is a thing.

      But yeah, parallel computing has taken a lot of the urgency away from that.Report

      • veronica d in reply to Oscar Gordon says:

        Well, you’re a fluids guy, right? So yeah, y’all need a lot of computing power. For simple n-body problems, particularly for uniform spherical bodies, it’s really trivial from a computation standpoint — unless N equals like 10^10 or something silly.

        I have an neat text on orbital dynamics that covers non-spherical bodies. It’s really interesting. It turns out that to make it computationally tractable, they introduce “gravitational moments” as corrections in the diffeqs, similar in spirit to the “dipole moments” and “multipole moments” used in electrodynamics. This really only works for “sphereoids,” where the moments are low degree. I doubt it would help much for small non-symmetric objects. Anyway, by adding moments, you can explain tidal locking and things like that.

        Regarding fully non-symmetric objects, it seems very few dynamics texts cover them, so naturally they are the most interesting.

        Anyway, diffeqs — cool stuff.Report

        • Oscar Gordon in reply to veronica d says:

          Yep, love my diffy-Qs. And yeah, my workstation that I use just for development and testing is 40 cores (not 20 cores hyper-threading, 40 actual cores).

          The actual work clusters for the big problems are orders of magnitude bigger.Report

          • veronica d in reply to Oscar Gordon says:

            That’s quite a few cores 🙂Report

            • Oscar Gordon in reply to veronica d says:

              Current project is modeling gas turbines, I can keep all 40 cores busy even on the simple cases.Report

              • DensityDuck in reply to Oscar Gordon says:

                We had a joint-design course recently where the instructor complained about everyone just throwing everything into FEM these days and increasing the mesh density to near-infinity.

                His contention was that doing this works, and for complicated joints it’s actually faster than working out the analytical solution, but you can’t look at a FEM and say “okay what happens if we use bigger bolts, more bolts, different placement, more or fewer washers…” the way that you can with an equation.Report

              • Oscar Gordon in reply to DensityDuck says:

                How long ago was that course?

                Past 10 years or so, we’ve gotten really good at incorporating Design of Experiments / Design Studies into simulations. Like, wicked fast plowing through design iterations where you can easily use bigger bolts, different placement, etc. (although in my field it’s something like swapping out a blade, or part of a wing, or an inlet, etc.).

                And part of that is because we’ve realized that the quality of the mesh can matter a lot more than the density. So sure, you can mesh down to the molecular level, or you can put better intelligence into your meshing algorithms and have it optimize the mesh.

                We even have algorithms that can optimize the mesh mid solution, so if the solution starts to show that an area with a dense mesh isn’t developing anything interesting, the algorithm will pause the iterations, and decrease the density in that area (and vice versa if something interesting is happening and more density is needed to resolve it).

                As for DOE, once the solvers got smarter CAD, and/or learned how to integrate with CAD tools, they could block up the CAD and the associated mesh so that when something is switched out, only that bit is re-meshed (if it wasn’t already meshed ahead of time and just waiting to be swapped in).

                One of the most powerful tools we introduced a few years back was overset mesh, where you would mesh a whole domain, then mesh a smaller domain around your part, and then you can have your part moving through the larger domain. We have a very cool simulation of two F1 cars (one overtaking the other) on a track, with each car having their own overset mesh, and modeling the airflow as they passed.Report

              • DensityDuck in reply to Oscar Gordon says:

                “How long ago was that course?”

                Last week?

                You’re not wrong that in practice it would probably be done by “pres butan to go plz” but if you want engineers instead of button-pushers then they need to understand what a joint is doing and what changes will make it work better, instead of spending two weeks optimizing fastener placement to sub-mil precision when the real answer is “don’t use titanium bolts”.Report

              • Oscar Gordon in reply to DensityDuck says:

                I guess it depends on what question you are trying to answer.

                First off, let me start by re-iterating that computers are really good at performing repetitive tasks, and can do that to very quickly solve problems and produce answers. And those answers can be very very right, or very very wrong. The computer doesn’t know or care.

                So if your approach to engineering is “pres butan to go plz”, and your engineers don’t understand the problem they are trying to solve, they are not going to get good answers to their problems.

                So for the engineers who complain about button pushers, I ask, is that truly how you see engineering being done today? Because, you know, this is my field, and very few engineers can be gainfully employed as button pushers for very long* (that’s what we have interns for!), so if that is what you think, perhaps you’ve been out of the trenches for too long. Additionally, I took quite a few computational physics courses (CFD/FEM/Validating models/etc.) in college, and the instructors in every single one drove home to fact that even the best models are only as good as their inputs, and the button pusher is obligated to understand when the output looks fishy and why.

                As for your original point – what is the question you are trying to answer? Are you trying to figure out how much load that complicated joint is under? Then FEM is probably the way to go, because the analytical solution would take too long to hammer out and validate. Once your have that FEM model, and a good idea of the stresses, then looking at the shear modulus of your standard bolt, and deciding Titanium is the way to go doesn’t need a FEM model.

                But, if you have no idea what kind of joint you need, or you are trying to find the strongest joint that lets you minimize cost or weight, then a FEM design study is a really good idea.

                *This was a big part of why I left the Lazy B, because junior engineers were being relegated to button pushers, and the senior engineers couldn’t be bothered to mentor anyone. The place is suffering serious brain drain.Report

              • veronica d in reply to Oscar Gordon says:

                Reading posts like this makes me wish I’d gone to university. Granted, I’ve done pretty well in software, despite being self taught. But still, I suspect I’d have made a really good mechanical engineer — or something like that. Maybe even fluids! It seems so deeply fascinating. However, unlike CompSci, it’s not really the sort of thing you can get into without university.Report

              • Oscar Gordon in reply to veronica d says:

                Sure you can, just study for and pass the PE exam – boom, you are an engineer!Report

              • veronica d in reply to Oscar Gordon says:

                Sounds easy 🙂Report

              • I guess I was just ahead of my time back in the 70s when the optimization prof would be handing back homework and complain, “And Mike, of course, simply beat the problem to death with a computer.”

                Other students would have preferred to beat the problem to death with a computer, but had not arranged for an unlimited (for the day) supply of CPU cycles.Report

              • Jaybird in reply to Michael Cain says:

                “Instead of carrying the rock, he put it in a cart and, get this, didn’t even pull the cart himself but had a horse do it!”

                The only problem is when you end up in the wrong “London”.Report

              • Michael Cain in reply to Jaybird says:

                Granted, part of the purpose of the class was to learn some of the clever ways to turn what looked like a hard problem into something much simpler. I learned that too, and got my first job at Bell Labs when, during part of the interview at the Labs, I applied such a trick and handed the interviewer a solution to a problem he had been struggling with for some months.

                But hours are always in short supply in graduate school, and sometimes taking 15 minutes to code up a brute force search and letting it run for a few hours is personally optimal. Especially since I had arranged a nearly (for then) limitless supply of machine time.Report

  8. Pinky says:

    ThTh1 – A throwaway example in the Summa Theologica (1265): “For the astronomer and the physicist both may prove the same conclusion: that the earth, for instance, is round: the astronomer by means of mathematics (i.e. abstracting from matter), but the physicist by means of matter itself.”

    Dante had his character in the Divine Comedy (1320) travel downward through Hell, and get disoriented once he passed through the middle of the earth, because gravity was pulling him in the wrong direction. They came out on the other side of the earth and remarked about how the sun was 12 hours off and the star patterns were different.Report

  9. JS says:

    [ThTh9] — Malaria is caused by a parasite, so this is a vaccine against parasites?

    Setting aside malaria itself (and if it’s even 50% effective for a year at a time, that’s twice as effective as the next best solution — bed nettings, so we’re talking a MASSIVE improvement in health and human happiness), that’s pretty amazing right there.

    So doubly ground breaking — hopefully a massive blow against a pernicious cause of much human misery, and a tool to fight a whole raft of other similar problems.Report

    • Michael Cain in reply to JS says:

      The parasite is a single-cell organism that has a life cycle with several forms. The form that is transmitted from mosquitos to humans is carried by bloodstream to the liver where it infects a specific type of cell. The vaccine leads to production of antibodies that attach to that particular form and blocks its ability to enter the liver cells, so it can’t produce the form that really messes us up.Report

  10. dhex says:

    [ThTh2] The weird (to me) resurgence of astrology as a thing that people talk about in public without fear of being laughed at feels bad. It’s especially popular with my wife’s students. I don’t know if it’s an ornament (like two people talking about a sports team or literature in passionate and nearly concrete terms) or genuine belief. The former, while aesthetically yucky, is far, far less worrisome than the latter.

    Odd note, unlike other anti-materialist conspiracy theories about the world (tradcaths, meyers-briggs, etc), astrology appears to be almost entirely consumed and propagated by women? I know a few dudes who are into astrology, but by that I mean two. Tradcaths are much more diverse (for a single sectarian group, I mean) and lord knows the I’m a TCYB ROFL (and this is why I’m a complete prat and you have to excuse it, as the sequencing tends to go) seems to be evenly distributed among men and women.Report

    • Pinky in reply to dhex says:

      It always had a place on the puzzles and comics page of newspapers & magazines. People have always talked about it in fun, or pretending like it’s only in fun. As for women, come on, why is it that women would be more likely to think about body changes based on planetary movement?Report

  11. Brandon Berg says:

    I think I’ve mentioned this before, but while I’ve heard the story about how Columbus incorrectly disputed the consensus view on the circumference of the Earth, I’ve never heard how his apparently successful return from the Indies affected the consensus.

    Did a bunch of people say, “Well, I’ll be. That crazy bastard was right after all!” and accept his estimate as correct, or did they immediately realize that he’d run into some previously unknown islands in the middle of the Atlantic Ocean?Report

    • J_A in reply to Brandon Berg says:

      Nope, it didn’t. Basically because there were land masses significantly south of where it was known the Asian land mass ended, they concluded, correctly and very quickly, that this was a totally different place. And they called it America, for the guy that showed this was nothing like Asia, Amerigo Vespucci.

      Sailors didn’t have a way to accurately measure longitude until the XVIII Century, but they’ve been measuring latitude very accurately for several centuries by then.Report

      • Brandon Berg in reply to J_A says:

        Yeah, but Vespucci didn’t do that until several years later, right? What happened in the intervening 5-10 years? On his first two voyages, Columbus just went to Cuba and Hispanolia, which lie roughly between Taiwan and the Philippines in terms of latitude. It wasn’t until his third voyage that he reached the South American mainland.Report

        • J_A in reply to Brandon Berg says:

          It’s unclear exactly who discovered or found out what and when. Vespucci claimed later in his life that he did understand they were exploring a New World by 1497/98 but many quasi contemporaries writing a couple of decades later believed Vespucci himself forged some documents later in his life to claim the primacy.of the “discovery” of the New World.

          But by 1501, when the coast of Brazil all the to Rio de Janeiro was explored it was very clear that this was not the Asia everyone knew about. Besides being in the wrong side of the equator, this land was significantly less inhabited and vastly less developed that what Asia was known to be, with animals and plants that were not what the explorers expected.

          Asia was far from Europe, but not unknown. People knew people that knew people that had been there. They had an expectation of what they would find in Asia, and that was nowhere to be seen (finding the Mexican valley and Andean civilizations were still decades in the future) Even Columbus claimed that he had only reached some islands west of Asia, like the Canary or the Azores islands were with respect to Europe, and not Asia itself.Report