wumpus

Which is extremely disappointing for an engine which at first appeared to be the holy grail. It sounds like it isn't doing variable displacement, which is probably needed to really get the (low load) efficiency up. But it really says a lot about ICE that they really can't beat the mileage of the original Prius. On the other hand, electric cars are hardly limited to sane levels of power (in ICE you can have power or efficiency [cheaply]. With electric, you can have both [and it will cost you]). "Hybridizing like a train" apparently is too expensive, although not terribly unlike what the Honda Accord hybrid uses. That has an ICE that either runs a generator or directly powers the wheels (without a transmission, but through an extremely "overdriven" differential). That sounds like an ideal setup for a hybrid, especially one that is more "plug in" (with significant battery range). Pure electrics seem to do surprisingly well, I have to wonder how much is leaving out the wildly complicated automatic transmission (and engine, but the transmission is more complex and probably more expensive). I'm sure Henry Ford had to explain why you needed to have something as explosive as gasoline in your "horseless carriage" as well. Seems to work out (in everything but the Pinto and some modern exotics. And don't ask about old magnesium Porsches. German firefighters learned to simply let them burn to the ground/road rather than try to put them out).

I remember hearing about "cruises to nowhere" decades ago. I'm guessing that this was when you had to go to Atlantic City to legally gamble or take the cruise (and cruising out of Baltimore beat going to Atlantic City). It was a thing, but I think it was a day trip (it only had to go a few miles off shore to open up the casinos). Part of the draw will always be that "I've been to space and you haven't". Legend claims that considerable bits of the Kama Sutra are impossible in a gravity field might also come into play. Anybody know how important the mile high club was to early aviation? Although I suspect that anything we are talking about should be seen closer to the barnstorming era (moreso for suborbital flights) which certainly didn't include anything like that.

Laser ignition is less efficient than compression-based ignition which Mazda is about to bring to market. Nissan made an engine that is pretty much "everything except ignition improvements" and I doubt it could break 40mpg. https://www.youtube.com/watch?v=A6H66xfEZC4 Compression ignition (like a diesel, only with normal gas fuel) should help ICE last a bit longer, but I suspect they will soon be largely used as range extenders for hybrid cars. Electric cars have too many advantages (except range, where a tiny ICE can help). Note that if the engine is small enough (like a prius engine) it is relatively easy to make extremely efficient at normal cruising speed (and let the electric be efficient at everything else). Flash memory is already going 3d, but logic is already limited by power density. Going 3d simply makes the power density even worse, although I'd love to see a two-layer CPU (or GPU) where one layer is standard logic and the other is a DRAM memory chip that is used as a cache with arbitrarily high bandwidth to the chip (DRAM and logic use essentially inverted processes, so the DRAM that went on your XBOX (or whichever chip used a lot of it) didn't have nearly the density of your normal "off the shelf" DRAM). There used to be a lot of research into things like "reversible computing" that would be far more efficient, but modern 21st century transistors simply leak too much power (a nano-amp here and a nano-amp there and with a few billion transistors it really adds up). I'd put internal combustion engines at the top of this list. There's a few more places to eke out a bit more efficiency, but electric engines are almost certain to sweep them into history before they can get there. The engines used by power companies to generate power might well also be on this list. Not thanks to upcoming obsolescence (it will take acceptance of nuclear power or an obscene amount of batteries to do that), but to simply hitting Carnot's efficiency (or really, really, close). You simply can nae get more power without breaking the laws of physics. Speaking of the "laws of physics", I've heard that error-correction-codes can pretty much hit mathematical limits. This doesn't mean that everybody that uses ECC will hit those limits (the stuff that does takes some time to do all the calculations), but for things like Starlink (woo-hoo, actual space reference there) expect to see near-Shannon rates (either through Turbo codes or LDPC [LDPC was a curious "forgotten tech" invented by then man who "wrote the book on data coding" but didn't include LDPC because it was a theoretical exersize and he was writing about important and real work. Fast forward a couple of decades and someone develops the Turbo codes (a development that was such a jump that it inspired claims of fraud. You know you have something patent worthy when big names are claiming it can't possibly be true). This lead to digging through the records and finding LDPC from said "guy who wrote the book"''s thesis (which showed a realizable system that could asymptotically approach ideal rates) and realized that modern computers could do the calculations. I'm not sold on batteries being on this list. Certainly, fuel cells or ideally liquid batteries (closed loop reversible fuel cells) would be better for most situations, but batteries always seem to improve by a fixed small percentage every year, and as we've seen form Moore's law that can really add up. I'd be really curious if methanol could be used as a car/truck fuel for a fuel cell (hydrogen simply seems too much of a problem in itself. The fact that it isn't overwhelmingly used in space (and not at all in aircraft) makes me laugh at the idea of using it in cars). When batteries stop steadily gaining in power density/cost/power you can stick them on this list. I don't think this will happen as long as a significant percentage of the worlds gross product is spent on cell phones and other battery operated equipment.

While the 80% solar areas are great, I wouldn't ignore the ability to not require any shielding from solar radiation (especially flares and similar). Even if Artemis can't do it, I would still expect any longer-term habitats to go "where the Sun don't shine".

I suspect it wouldn't have taken all that many men in Kuwait in 1990 (assuming you got them there fast enough) to avoid a (and in hindsight, two) war. Granted, I'm pretty sure the US had a lot of forces in Saudi and still didn't invade Kuwait simultaneously with Iraq, but with a sufficiently rapidly deployable force it might have been an option.

I remember some program to do something similar at the end of the Cold War. The idea was to drop two (might have been more than platoons, probably less than two divisions) groups to anywhere on Earth. Time to move it all (including gear): 48 hours. Firstest with the mostest isn't likely to change, but how fast you can get there first has.

That and transistors are as close to free as possible. In CPUs, the limiting factor is likely clockspeed and the how complicated an instruction is (unless it needs to interact with memory) really isn't an issue. Once vectors get to 512 bits long or so, power becomes a problem, but otherwise ungodly complicated instructions "just happen" every cycle. It might make sense to optimize GPU instructions to use something more simple than IEEE-754, but current GPU companies won't be interested as long as AI developers (and other non-graphics GPU buyers) want something based on floating points (Intel and cell phone GPU manufacturers might be more interested in optimizing graphics operations).

Note that only "floats" are unbelievably fast on consumer GPUs, doubles typically are greatly reduced (thanks to market segmentation, often much worse than simple transistor savings would imply. And equally often by disabling working double logic where present for the "big boys"). CPUs tend to focus more on doubles, although equally capable of cranking out as many floats as they have register ports (for those bits). I'm not even sure they need to pipeline anymore (although I expect if you dig deep enough, plenty of pipelining has to happen even if they can execute back-to-back instructions). Fused floating point multiply add is also pretty common. Believe it or not, the biggest stumbling block is if you need to round on the intermediary multiply. IEEE-754 appears to make floating point as hard as possible to actually lay out in transistors (example: you have to compute the full 112 bit multiply and then round to the exact 56 bit result...), but nowadays they seem to not have any problem including the full suite on GPUs. Divide is still a pain. Any pipelined or parallel algorithm is going to have nasty limitations such as inverting the thing then multiplying [favored in early vector extensions, I suspect it might be standard practice today] or converting back and forth from fourier transforms (which greatly optimises multiplication of >>100 bit factors and makes division as easy as multiplication). x86 will have to update each flag for all flagged conditions the same as 8 byte and 64 bit int operations (assuming using an ALU and not the SSE/AVX interface). This means the internal datapaths will be a lot more complicated than naively assumed (although the programmer will never see the difference). I'd assume other architectures have similar things that were assumed trivial whenever they were designed. I'd assume that any log operation would be based on base 2, not base e. This should allow for much more producable algorithms. This sort of thing would be ideal for things like graphics and audio, mostly because our senses send more-or-less log-based signals to the brain (or at least we interpret such signals logarithmically). Don't expect any great gains unless you want the output in log form. You can more or less effectively create a nth order differential equation simulator out of opamps and design in ways to change the coefficients on the fly. But they'll never be "generally programmable". And the "butterfly effect" becomes pretty clear long before that: consider why we even have double precision in computer. A "float" is good for seven decimal places of accuracy. A "double" is good for ~16. Only the most rigorous physical experiments will get 7 decimals of accuracy, and nobody is getting 16. But it isn't just the "butterfly effect": rounding errors (even with "perfect rounding" griped about above) will eventually corrupt even the most linear and stable algorithms. I remember doing a 32k Fourier transform on audio (16 bits, about 5 decimal digits) using just floats and the audio output quality was worse than 8 bit. Building a mechanical or analog computer to 70dB (a float's accuracy) of accuracy (each operation, so expect things to get degraded each operation) sounds possible. Getting to 150dB (double) is likely impossible (and no, the only way to break it down into "hard" and "simple" is by going digital).

I'm curious what the specific case is now. I know that after 9/11 access to Goddard Space Flight Center is pretty restricted and employees/contractors need some sort of security clearance (although possibly some type that allows non-US citizens: more about not shooting the place up than espionage). There isn't a whole lot of difference between the access control of NASA Goddard and Naval Research Labs (at least from a visitor's standpoint). I'm sure when the contracts are handed out, anything classified means the DoD gets to defend their turf. But I suspect NASA is allowed to keep quiet about a bunch of stuff, just nothing at "secret" or above.

I doubt air breathing is anywhere near the next step. I thought it was, then space-x decides to try to catch fairings. Cost of a fairing: $6MBucks Cost of fuel: $0.2MBucks Even occasionally reusing a fairing is going to save more money than if they were magically gifted the ability to use seawater instead of fuel & oxidizer (although obviously using less fuel means less complexity, although nearly all this mass savings would be mostly limited to the first stage). Obviously, low fuel will someday become important, but how unimportant it is (especially when you already have working inefficient systems. Competitors might think about higher Isp rockets. Unfortunately one of them was Stratolaunch who completely got the idea wrong).

About four hours before your post: Scott points out that NASA typically takes anything that needs space testing up to the ISS. But for anything the DoD doesn't want a cosmonaut waving a cell phone around the thing and taking all sorts of measurements, there's X-37B.

I'm not sure being a persistence hunter is all that different from being a migratory species. And most of the reason we aren't a migratory species is simply too many humans to allow that (humans are too territorial to allow *other* migrant humans to waltz through). It is entirely possible that humans were pretty migratory back when evolutionary forces on human strength levels stabilized. Granted, I wouldn't expect anything intercontinental: during such huge migrations somebody would get the bright idea of fortifying obvious passes and demanding tolls from migrants. Human internal carbohydrate sources are good for running ~14-18 miles (this becomes painfully obvious while running a marathon). Expect to walk large parts of longer distances, all the while consuming food (so humans would have to forage while migrating). Backpackers likely go much further than this, but there is a reason that when you think of nomadic tribes you think of people on horseback (I suspect most of the horses are pack animals).

Is Space-x still planning on open-loop "sweating" cooling in Starship? Pump fuel/oxidizer (whichever has the best cooling capacity) out through tiny holes and use that for active/ablative cooling? And I'm not sure what else you could mean by "active cooling": it isn't like there is an available source of cold atmosphere to flow through a radiator. You could pump heat to superheat a radiator (which could black-body radiate if you had to), and maybe by the time flowed around the spacecraft it will have re-expanded and cooled off (so it *could* flow through a radiator), but these all seem unlikely.

True, but you still need an unimaginable launch cadence to get to the point that any gains in lowering vehicle complexity outweigh the additional launcher complexity. Adding fuel and thrust is pretty easy (adding solid boosters to non-crewed vehicles is even easier for increased thrust, then all you have to do is add additional fuel). Once you are at the point where a somewhat supersonic rocket is relatively easily recoverable, the only question comes down to is the launcher less complex than the rocket (doubtful, although you might want to leave the atmosphere for staging) and which uses less fuel. I'm also curious just how much benefit would be gained by merely moving the vertically launched rocket to the maglifter launch site. That would almost certainly built at elevation, and the lower air pressure would bring a lot of gains. Strap on some solid boosters to get your TWR up, and you have a similar system without bothering with building the whole rail assembly (and issues of getting the rocket structure to support its full weight in multiple directions).

Would only really make sense for a Concorde replacement. Ramjets might be a plausible replacement for jets, high bypass turbofans not so much. Back on topic, before asking this forum about something fairly well studied (ground-based launch schemes) you might try googling our patron saint: Considering Elon Musk came up with the whole hyperloop idea and then gave it away (presumably because he couldn't afford to develop it, nor was worried about people using it to compete with him), I don't see why you would think that he never considered using it to launch rockets. The biggest issue for this type of thing is simple: fuel doesn't matter. It is hard for anybody new to rocket science to get through their head (and probably doesn't help any sci-fi work to get it right), but until rocket science becomes incredibly more mature, there's no reason to optimize for fuel. Fuel means mass, and optimizing for mass makes sense, but real rocket science isn't like KSP and fuel tanks are far, far lighter than you might think. We can typically afford to lift the fuel. Last I heard, Spacex charges ~$60-100Megabucks to launch a rocket. Fuel costs are $0.1Megabucks. Until you get rocket re-use to be a trivial thing and make launching a rocket almost as cheap as launching a plane (some of the people cheering in the background of spacex launches are presumably non-launch employees, but plenty of them are certainly doing required work). And spacex is only used as they are both far more transparent to their competition and tend to have reduce costs elsewhere more aggressively. It will take some time before anyone cares enough to try to reduce the cost of fuel. Long ago I computed a rule of thumb that said that a rocket basically consumed half it's mass in fuel for every "mach 3's" worth of velocity. I think it was based on the Saturn V (with a TWR of ~1.2) and I doubt that it true of a Falcon (or similar) that can take off with a TWR of ~1.5. Even so, the amount of design needed to add ramjets or scramjets to a rocket to get it to mach 3-9 would be formidable. I could see such a thing [possibly just some sort of "starship heavy" with air-augmented boosters] for a "starship refresh" (of 2030-2040), but no earlier. And even then I suspect that there would be plenty of lower-lying fruit before they would bother with trying to reduce fuel consumption. PS: I heard that Blue Origin scooped up DC-X team members early on. One thing that DC-X is said to have clearly demonstrated is reduced launch costs. If Blue Origin can get something like New Glen (or New Armstrong) out the door and into orbit, they might be far enough along the "low lying fruit" path to proceed to fuel efficiency...

1.8 is probably "good enough" for KSP. Perhaps unity will make a good raytrace mode someday, but I suspect 1.8 will be fine. Consider Sid Meir's Civilization. There are plenty of people who will prefer any of the games from 1 to 6. I'm partial to 4, primarily thanks to Leonard Nimoy's voice work. You'd think Civ 1 or 2 (1991 and 1996) would have few players, but checking the civfanatics website, people are still commenting on those games' fora. Certainly the number of people playing KSP 1.0 will be greatly diminished, with very few new players. I'd go so far as to say the pre-release game might as well be a different game: but I'm sure there are those who still play it, but very few and the modding scene is limited. The important thing is that Squad never implemented Steam's DRM (or any other DRM) on KSP. You can run KSP no matter what Squad, Take 2, or Valve do to their servers. So the old chestnut of "the developers burn down all the old editions" is clearly not true for KSP (although it might be true for more multiplayer oriented games).

I'd recommend cast iron or steel. With titanium always the danger of shattering the keyboard, and with super strength you wouldn't care how heavy it is. And super strength had better come with "super indestructibility", or you'd have all the injury issues of your typical 7 foot NBA prospect, although I think we've covered this above.

True, but I had to ask myself "how would your life change with super strength?". And the answer would be "as much as it would had I been a competitive weight lifter or similar". One exception might be those who live on a low-gravity (or possibly with a very dense atmosphere) planet and flying (via some sort of wings attached to your arms) was a common means of travel. Even then, it seems like something that would just add color to your sci-fi (think RAH's "The Menace from Earth").

Humans (and wolves) favor persistence hunting (simply running the prey to exhaustion, and then easily taking it down). I'm not sure what the wolves use, but humans have great endurance thanks to sweat glands. I suspect that slow-twitch muscle fiber also plays a role (and is far less present in gorillas and similar). It is unclear how this could evolve on its own (especially since it appears likely that the reverse happened, we used to be much more similar to chimpanzees and presumably had similar strength). "Super strength" isn't what pushed humans to the top of the food chain, and I'd be surprised if neanderthals weren't considerably stronger than us. Perhaps some sexual selection system where jocks would have much greater reproductive success than anybody else (presumably great hunters always had such advantages, but shear strength was never the main key). A quick look at the results of a powerlifting championship (or perhaps watching some [American] football) would show that there are those with "superstrength" (or multiple times normal human strength) walking among us. It doesn't seem to have dramatic life changes outside of athletic competition.

The title of this thread includes "refueling on Earthlike worlds". There are two "Earthlike worlds" remotely within reach: Mars and Venus. Refueling on Mars is a known issue (details of difficulty are of course unknown), refueling on Venus would be pointless. Worlds around other star systems are so difficult to reach and would require such unknown engines to make the thread pointless. Blocky/non-blocky entirely depends on the standard size of material available vs. size of the hull. Since steel is made in sheets much smaller than the size of a hull, nobody ever considers making a blocky ship: it would cost just as much (well more steel means more materials cost and more welds so it would cost even more) and be less effective in everyway. Something the size of a car can have each panel stamped (hydroformed) to shape (instead of being welded from little panels) and thus are a bit more blocky than might be optimal for their function. Racecars and general aviation planes might be of similar size, but often are built as a monocoque for strength-weight efficiency (but harder to manufacture).

They managed not to say a word about SLS (not so about NASA) when Marshall's Science and Technology Office manager decided to present a paper to the American Institute of Aeronautics and Astronautics Propulsion Energy Forum and Exposition on his new reactionless drive. https://arstechnica.com/science/2019/10/corkscrewing-bouncy-ion-drive-would-provide-thrust-in-different-universe/

You still need ~3000m/s delta-v from LEO to the Moon. You can either boost mostly rocket up from sea level or boost mostly cargo + ion power and wait much longer. And the same is true for Mars. I'd assume you would lift the trans-Mars habitat into Lunar orbit and then use gravity tricks to move it into a more elliptical orbit. Then send the crew and the final 1000m/s boost to Mars. If you want to get to the Moon in 2025 (Congress appears to have scuttled 2024), forget the whole idea of ions. But if you want to do it with less launches and cost, it should definitely be on the table.

LEO used to be halfway to anywhere (RAH). If you are willing to wait until ion-powered spacecraft maneuver fuel-depots into position it is 90%+ of the way anywhere.

It should be feasible to compute how many man-hours went into crafting plate armor (although the number of different specialties would make it very time consuming to learn enough to do it): time to mine the ore, time to harvest the charcoal, time to smelt the ore into steel, time to craft the armor (presumably a much more elite craftsman than typical). Also depending on the time and era for "medieval" there might well be records of slave sales: slaves were often the most valuable things traded on a wide scale (the Normans [and I guess Franks] didn't like slavery, so it disappeared in England somewhat after 1066 and I assume had already left France). Another way to see just how valuable human labor was (to the elites) was the effects of the Black Plague. Afterwards everybody knew just how valuable commoner labor was to the elites. And there was always the longbow (see "Pure Newtonian space combat" thread). Archers didn't require all that much valuable equipment (at least until people noticed yew trees disappearing), but once trained they were invaluable. Same for pikemen (although I'm not sure how much was training and how much was selecting extreme bravery. Holding that pike during a cavalry onslaught must have been terrifying). While elites may have talked a lot about "land", what they meant had to include the people they held in disdain or it wouldn't have been all that valuable ("owning" a mountain range would be pretty pointless, aside from strategic benefits).

While Mad Jack didn't make his men fight with longbows and broadsword, Ben Franklin was pretty serious about outfitting troops with longbows during the [US] Revolution. I'd assume that issuing half the redcoats large shields (one holds the shield, the other fires a musket. Both hide behind the shield) would simply crush any archery troops. Of course, that meant they would have to carry the shields around while fighting the standard musket troops just in case they met archers. Probably the real reason it didn't fly was the difficulty in training longbow troops (which was just as much a problem in the hundred years war, except they didn't have muskets so the longbowmen were trained). Both Cortés and Pizarro won under those conditions. Granted, they had steel armor to keep the arrows off. I've heard that Little Big Horn was fought before the US Army had entirely converted to brass cased ammunition and that the earlier ammunition tended to jam a lot. Not that Lakota tactics didn't help a lot as well. The Zulu case is pretty weird: the primary reason you want a breachloader is to avoid enemy fire (you can lie on the ground and reload. With a muzzle-loader you have to stand there as an easy target), while this really isn't nearly the advantage against Zulu weapons. If you bring a musket to a Zulu war, you better do it more like a conquistador (be ready to use earlier weapons as well) as the Zulu is going to go from outside of musket range (>>100 yards) to spear & shield range before you can reload (not true of muzzle loading rifles, but you have to be a good shot to make that second or more shot count). The Zulu lost an entire generation in that one battle (the army was *all* male teenage Zulus, so losing a significant fraction of the army was the same as losing that fraction of a generation). Nevermind King Pyrrhus's lament that "another such victory and we are ruined", that one victory "broke their heart" and essentially lost the war. Victorian England barely noticed the loss (and PR mostly concentrated on Rorke's Drift).