The patent for a supercloud server

A patent now exists for the heliosync orbit: US-12679564-B2 at ppubs.uspto.gov, for datacentres which I'd written off two years ago as "cockamamie". Various factors as of 2026 are making this less silly. Sophia Space has press-releases, linked at Zim's place.

Heliosync in theory runs 400-1000 km so, LEO. In practice they don't want drag: superfast atmo, which degrades... everything. They also don't want the congestion which reaches further, maybe to 600 km. So they're thinking 800 km. They must beam the data to a comm array anyway, which may as well be SpaceX' constellation of Starlink v2.

Heliosync around Earth gets about 1.361 kW/m² "flux", more than that most the year because of eccentricity. The patent illo assumes they can drive 1 kW/m² from the solar-cells, which they expect η=.3. They leave the material to others; maybe Perovskite. Anyway I read 400 W/m² available to the computer, not 1000. Also here are transmitters, of data.

Some of that flux is reflected (5% here); less is transmitted. Most that heat needs radiating-off, still maybe 1.3 kW/m². The good news, if they insulate the computer: the computer is wide, flat, and narrow. They think that the shaded darkness will allow "passive" radiation through the shade, into space. More good news: most office CPUs seem to run in the 45-65 W range as long as you are not gaming on them. 400 W/m² might make for a respectable server-farm.

The design is simple if they can stow the volume onto a rocket chassis. Direct 800 km looks like a Starship job.

In addition, if the Lunar south pole gets a superstructure, similar architecture might work there too. For a start. (Later we expect coolant-cycles underground.)

Loanwords in Choltian

In pictographic old Egyptian, loanwords are easy to suss. Not having any indigenous roots, such words are spelt phonetically and never otherwise. Magnus Pharao Hansen has been applying the same principle to classic Choltal inscriptions.

They have a loanglyph too: the Striking Owl. This iconography is Teotihuacano. I think everyone agrees that the Place Of Reeds knocked a lot of head down in the Petén. The empire even installed a prince to rule over Mutul ("Tikal"). But we're talking about intrusive words that haven't a contemporary Maya-family etymology.

This'll be a placeholder. (I am not a mayanist.)

Dr Hansen has been arguing that the loans are in some paraNahua dialect and that, therefore, the Place Of Reeds was, if not called anything like "Teo Huacan" at the time, also paraNahua. Mind: this depends on his ability to reconstruct what that dialect would have sounded like, a thousand years before Catholics came with a Latin alphabet. We can do it for Choltal, precisely because sometimes an author could get phonetic. But not many outsiders were using so good a system, then, for writing.

That central Vale was shared between Nahua and Otomí as of AD ~1500. It may be some sharing had happened already. One does wonder how the Purempecha interposed themselves into Michoacan in the meantime.

Anyway if you want to give Dr Hansen a hearing, try "What Happened When?: Relative Chronology and the state of pre-Nahuan in the Classic Period. Paper presented at the Relative Chronology workshop at the University of Copenhagen, June 30th, 2023".

n-body, near O(n)

After Newton discovered the equations behind Kepler's 2-Body solution, he ran afoul of the n-Body Problem. A limited solution presented itself in the equilibrium point; the full five solutions were offered by Lagrange. Further refinements handled special cases of the 4-body like when Laplace sussed out three of the heaviest (and largest) Jovian moons. So here we are in the computer age, when computers struggle to handle particles numbering more than maybe five. It's O(n*n).

Kepler and Newton assumed that the bodies are rigid points and that we ignore Einstein; so Io's tides and Mercury's depth in the well don't figure. Also computers have to approximate floating point numbers well above Zeno's quantum level (Planck?). That means we have already sold the horse for approximations. And that means we may as well approximate some more. Keyframe Codes has a little primer on the available algos.

Midlevel nerds have been using Barnes-Hut. This method runs on multipoles: it buckets the particles in squares-of-particles, finds the center of gravity, and runs the calc for the centers. The buckets are identified with quad-trees; Barnes-Hut ends up O(n * ln(n)). Leave aside for now if we want to run this 2D or 3D; we'll pretend perfect ecliptic. What if the cluster doesn't behave like a square (or cube). Real astrodynamicists need the particle swarm where most particles would be clustering. They will be looking first at Lagrange's L4 and L5 "Trojan" haloes especially for Jovians. These haloes are elongates.

In 1987 Leslie Greengard came up with something else, the "Fast Multipole Method" - usually FMM. Not only does it claim O(n) but, depending on how often you want to run each n, the error can arbitrarily diminish. This was Greengard's thesis and, it seems, got a refinement under one Vladimir Rokhlin. Discontents exist: Srinivas Aluru in "Greengard’s N-body Algorithm is not order N" (pdf) showed that the choice of precision must depend on n (at least ln n), therefore cf the title - to the tune of O(n ln²n). That extra ln(n) is, then, the price for accuracy Barnes-Hut must pay in the Trojan haloes. Dr Greengard's peers seem to have agreed; he got to keep his PhD and the IEEE considers FMM in the top ten of twentieth-century algos.

This starts by keeping each interaction as a field vector Φ and treating that vector as a number on the complex plane. The vector acting upon particle "i" is the gravitational parameter of counterpart "j", times the natural logarithm of the i-j distance (plus iθ); that's the force. The iθ exists, here, to make it a vector as to place the particle on the coördinate plane; a computer may ignore it now to figure the coörds when it's display-time.

Some might then use "numerical methods", which is mathematician for iteration. The FMM, in the complex plane, may use the Cauchy-Riemann equations. At this point I'm leaving the rest to the nerds.

The scouring of the land

I am continuing my scan of Elitzur's dataset. Yesterday I added some words I recognise in Arabic: BKY, ḌḤK; this morn have I removed his first example elón. I got intrigued with example #5 GLY so have held off to poast that, today, separately.

The Quran will use GLY for exile. The Arabs didn't, at least not in the Safa. There it meant cut off, erased, exposed. So when I look at Elitzur's data, he turns up Judges 18:30 for which authenticity he argues: which I think vocalised some gelút há-aretz in CBH construct. Clearly you can't exile the land. Meanwhile GLY as exile comes nowhere in the Hexateuch (to Joshua), despite exile being a central theme.

By parallel with Safaitic, gelút há-aretz implies a scouring. Elitzur applies this term to the land around Judges 18:30-1, whose candidate is Shiloh's temple which, by then, was dismantled. Whenever Judges thought that was: Gath's invasion? Hazael?

At this point I want to bring example #1, elón; I further suspect the há- as late in Hebrew's evolution (it was probably *han- prebiblically). Elitzur finds elón in Genesis and Deuteronomy, then through to 1 Samuel. Once David takes control we no longer read the word. For Elitzur it should be something like the Archaic Greek chora (district/ viable-flatland)... like aretz. Why does it end with Samuel, if it was Deuteronomic, he asks. I counter that everywhere, elon had been fossilised - into placenames (construct, without há-!) and/or semantically become "oak grove". These elón texts existed to place scenes into a landscape of memory. For Judges 18, either Elitzur must abandon its status as an ancient account or else he must admit that elón was already a fossil.

For my part I accept that Judges 18 be early. As for why the Pent-to-Hexateuch didn't use GLY, Elitzur would put that down to not knowing Judges 18. For Elitzur the Hexateuch is simply earlier. The Deuteronomic History, having read Judges 18, shifted GLY's semantics to exile. Elitzur fn. 17 is aware of GLY for exile in Akkadian (which might have followed a parallel shift) then Aramaic.

I won't argue that point for Deuteronomy alone. For the rest of Torah (and Joshua) I propose, rather: refusing this use of GLY. The authors and editors saw the term as belonging to a later time.

REPOST: I don't know WTF happened to the URL, but let's try to fix it

The end of the bombardment

We haven't discussed the "Late Heavy Bombardment" on this blog. As of 2024 we've rather been skeptical. We may, however, have some timeline for the end of some kind of bombardment: Universe Today reports from Tim[othy] Johnson. And from Hamed Gamaleldien and some others.

The issue here is that the Hadean has, in fact, ended - but the crusts came later than that ending, agreed 4.03 Gya. Crust first formed 3.9 Gya. There should have been, what, 400?500? My after the great Theia crash and all its debris; Earth should have cooled enough by then. Johnson argues that the crust had been heated exogenously. That is, Theia wasn't the last crash.

The "L" part of the LHB implies a surge at the 4.03-3.9 Gya ... I guess we can call it a "period", 130 My as it would be. We've been marking this period as the Archaean beginning. Johnson doesn't need a surge. A continual pounding from our uncaring universe would do the job just as well.

The constant heat over this extent of time continued the process of differentiation, at least at the crustal layers. We have a mostly "felsic" land surface up here, dominated by silicon and aluminium and all those lithophiles... just like the Lunar regolith. Granite looks compositionally felsic. The "mafic" rock, of iron and its siderophiles, clusters in some rare upwellings and on the ocean floor. That's básalt or should we say défault.

Our Hadean is defined by committee, basically: Ten oldest zircon U-Pb ages in the Acasta Gneiss up in Canada's NW Territory. Gneiss is metamorph granite so it was a crust once, exposed by glacier. These zircons are however preserving siderophilic (heavy) metals. So I dunno. They might be mafic and precede our crust - and the crust of Acasta.

Johnson and Gamaleldien earlier had evidence for water in those transCanadian zircons, which they haven't disavowed. So I guess they can continue to have a waterless Hadean, just like in the Gospel of Luke. Archaean is, then, defined by still-no-crust but now with uneasy pools of water. Maybe a steam period.

The implosive fusion pulse rocket

I woke up today with ToughSF poasting Gradient Field Imploding Liner Fusion Propulsion System. This report is fairly recent, like 2018. I wasn't quite blogging Atomic Rockets content then but, soon.

The "field" here means magnets. The liner is of some metal that can crumple when subjected to the Tesla. This implode-the-liner-with-dekamagnets tech is not new. Here is something from 1976. Also given the heat of the fusion around there, they cannot use superconductors - not 1976 niobium, not 1986 ceramics, nothin' under 125 K. Which puts MR Lapointe at some engineering conundra. For instance he only has 30 Tesla to play with, which he's wrapping externally. 1 T, internal.

Luckily for him and his, "magneto-inertial" designs had been an active, uh, field of research up to 2018. And he doesn' t have to keep the plasma: we're going to spaaace. The drive works by pulse like an Orion; unlike Orion they have to start in space.

This is about where nerds start talking z-pinch (or ζ-). This report is talking θ-pinch.

To load up these interplanetary missions, the 2018 paper suggests a couple of SLS boosts to eject 260 "mt" off our well. "mt" looks like "metric tonnes" so I'd say "T" or "Mg". As of last Friday Elon can sell you a Starship with throwaway booster (supposedly 188 T); he might soon be able to rent you one as well (for 115 T). They'd allow assemblage in orbit, some of those T being sufficient fuel and orbiting Starship to push your fusion rocket out of Earth's hair.

Where we speak of specific-impulse (I_sp) as usual the time-savings ramp up beyond Mars, like this takes 50-100 "mt" to Saturn in 200 days. In this context, 20 mt is a crewed mission. The report requires for the Earth-to-Saturn mission no less than 20000 s of I_sp, which I take as a function of deltaV more than of distance.

I_sp depends on what to use for the crumpling liner. Table 5 lists 32,200 s for Lithium (7.8 kN thrust), 17,145 s for Beryllium (14.45 kN); compare Sunbird's 15ks, Howe 5ks but 100kN. So lithium is how we, Earthlings and Venereans, get the direct lines to Saturn and beyond. Definitely going to have to construct the former in some anhydrous and anoxic environment like, uh, in space. Table 5 further calculates the energy required, for Mars: lithium to expend 165 MJ.

Asimov-approved beryllium would appear to work best where we don't care about longterm I_sp. A three-month trip to Mars ain't that much better than a six-month. Here we're optimising for thrust and, it seems, for energy (153 MJ). Energy aside that looks better for anything as far as Jupiter; inner-planetary travel or Oberth toward the Sun will give us lots of that energy, as well. With the extra thrust comes more payload and, we must add, more payload affords more options for crew comfort. Like spinning the hab, protection from radiation, all of it.

Also, pipe dream here, if we had stations on Vesta or on Deimos, beryllium could probably hit up Saturn thence (and it might even be mineable on that differentiated silicon). Launch windows will be a bear out there however.

Age of Mangione

Sunnyright, who got linked by Stephen Green, points to a 60 Minutes ep where a company invested $3B to develop a drug, which company now has to recoup its expenses. Commenters, who Deserve treatment, are calling for Luigi. They mean, for an army of DavidsLuigi Mangiones.

Green assumes all these commenters be Left. They are not.

A few days ago Ace of Spades posted this juxtaposition:

I'm just going to say it: I am now ready, willing, and eager to begin inflicting punishments on communists for what I consider wrongthink. If they want to criminalize my thoughts, I am now ready to use state power to criminalize theirs.

Mike Lee wants to defund schools that insanely continue to force students to get vaccinated in contradiction of their religious beliefs.

Ace has been so "ready" for some time longer than just this past week, I suspect. This whole side of X has been sliding into this, yes, fascist direction.

Meanwhile Richard Hanania wonders how come Bill Gates Jr hasn't st[oo]d up for himself and for public health, generally. I might so wonder, too.