Apollo and Artemis

Gene Wolfe’s Book of the New Sun is a great science-fiction/fantasy epic about a far, far future version of the Earth, where there were many incredible technologies developed, but they’ve all been lost. In the time of the story, people live in the decayed remains of a high-tech world, but understand very little of it, and get by using advanced devices that they essentially understand as magic. There’s a very well-known passage

The picture he was cleaning showed an armored figure standing in a desolate landscape. It had no weapon, but held a staff bearing a strange, stiff banner. The visor of this figure’s helmet was entirely of gold, without eye slits or ventilation; in its polished surface, the deathly desert could be seen in reflection, and nothing more.

This warrior of a dead world affect me deeply, though I could not say why or even just what emotion it was I felt

The picture being described is this one, or one like it:

The part of the Book of the New Sun that contains that quote was published in 1980, during the not-quite nine years between Apollo 17 and the maiden voyage of Columbia, the first Space Shuttle. At that time, the notion that spaceflight would be a lost technology so forgotten that a regular person couldn’t properly interpret a picture of a person standing on the moon was fanciful.

Today though, most of the Apollo program is lost technology. The F-1 and J-2 rocket motors that took Apollo to the moon cannot be built today, because no one knows how. Every now and then, someone floats the idea of building a team of technological historians to scour the archives and conduct interviews with elderly engineers in nursing homes to try and resuscitate the secret of these rocket engines, but it’s never been done. Some people believe that there was a certain degree of tinkering and know-how to those engines that couldn’t be learned anyways. It’s also been pointed out that much of the complexity of the F-1 had to do with the fact that its designers didn’t have digital sensors and electrically actuated valves and such, so they had to use a bewilderingly elaborate fluid-based system to safely turn on the engine and spin up the turbopumps and do all the other things in the exact right order. People have suggested that a much simplified version could be built using modern gadgets, but that’s never happened either. For practical purposes, the technology of Apollo has been lost.

After Apollo, NASA moved on to the Space Shuttle program. STS was ultimately a fatally flawed concept, but was as technologically daring as you can imagine. Instead of expending the rocket, the idea was to have a large orbiter (the familiar plane-like part that sat on the side at launch) that would carry the expensive engines all the way to orbit, then back to earth. The solid rocket boosters on the sides were supposed to be recoverable and refurbishable, with only the external tank (the big orange central component) being fully disposable. There were a bunch of engineering challenges, but two are worth noting. First, the Space Shuttle Main Engines (SSMEs) that were mounted on the back of the orbiter needed to fire from launch all the way through reaching orbit. On most rockets, there is a hard trade-off between having a rocket that is efficient in space (which requires a big nozzle) and having one that is safe to fire at sea level (where, for complicated reasons, having a big nozzle can destroy the rocket). The SSME does both by having a sort of virtual nozzle inside the main nozzle that allows the engine to fire safely at sea level, but then use a big nozzle in vacuum. In addition, (I believe) that the orbiters are the largest objects ever to safely reenter the atmosphere, and could not be built in the standard capsule style used for Apollo and other manned spaceflight missions. Instead they had an unbelievably complicated system of ceramic tiles to protect the orbiter from the heat of reentry. Developing the shuttle took about 8 years and cost perhaps $40 billion.

After shuttle was cancelled, Bush II proposed a program called Constellation, that would adapt Apollo and Shuttle technologies for a Mars-capable rocket. The program was over-budget and unsuccessful and was cancelled when Obama came into office. A cut back version of the program called the Space Launch System carried forward starting in 2010. The basic idea of SLS is that you take a Space Shuttle that’s ready to launch, and you simplify it by moving the SSMEs from the back of the orbiter to the bottom of the fuel tank. The boosters stay in the same place, but they stretch them out a bit. Then you get rid of the whole orbiter thing, and just put a second stage and an Apollo-style capsule on top. (The second stage is essentially a 50-year old design borrowed from the Delta II rocket.) An incredible fact is that SLS will actually be using the same physical SSMEs that went into space on Shuttle. They were, after all, designed for reuse and came back with the orbiters. Unlike Shuttle, SLS is fully expendable. The basic idea is to give up on Shuttle’s ambitious and troubled goal of reusability and kludge together a system using Shuttle spare parts that can fly. Not very elegant, but quick and dirty and cheap.

Except: SLS and its capsule (called Orion) have already cost $30 billion. SLS has never flown, and it was recently announced that it will not be able to fly prior to late 2021 (i.e., 11 years from the start of the quick/dirty/cheap design process, about the entire lifetime of the Apollo program, from conception through the end of operations). Even that date is purportedly not possible without a bunch more cash. Here’s a good thing on SLS from Oliver Morton’s new book on the moon:

(Mostly included for the line about Easter Island statues.)

Appropriately, SLS is the intended workhorse of the Artemis program, the parody space program that has been announced to commemorate the 50th anniversary of Apollo 11 and to flatter Trump’s ego with the idea that his administration could put people on the moon by the end of his (bite your tongue) second term. Artemis is supposed to cost another $20-30 billion.

There is a good chance (I would say >50%) that SLS never flies. Needless to say, Artemis will not put a man on the moon by 2024, and likely never will. The technology to fly to the moon has been lost. There is something very Wolfean about 2019 humans rummaging around in the warehouse, pulling out some mothballed SSMEs, and then trying feebly to cobble together a working rocket around them. (And take longer than it took to design and build the SSMEs AND the STS orbiter AND the rest of the shuttle system.) Forget using them for their intended purpose of reuse. Never even think of trying to build something better. Instead, burn up the physical legacy of the technologically sophisticated past to provide some sad light for the present. And, perhaps, fail to do even that.

I find lost technology to be a very enlightening way to view the world. It’s hard to see, because an exponential explosion in the power of digital computers has papered over the regression. Here’s a good piece about how modern software is horribly designed, but is able to modestly outperform older systems due to having many orders of magnitude more processing power at its disposal:

Only in software, it’s fine if a program runs at 1% or even 0.01% of the possible performance. Everybody just seems to be ok with it. People are often even proud about how inefficient it is, as in “why should we worry, computers are fast enough”…

You’ve probably heard this mantra: “Programmer time is more expensive than computer time.” What it means basically is that we’re wasting computers at an unprecedented scale. Would you buy a car if it eats 100 liters per 100 kilometers? How about 1000 liters? With computers, we do that all the time.

Or this piece about how scientific advances have slowed, despite exponential growth in resources (plus all those computers!).

The picture becomes clearer when you think about the physical world, where digital computers can’t carry the load for us. For example, the US has about 775 miles of subway currently. For 2019, according to this source, there is an additional… 1.7 miles of underground rail on tap for 2019. Highway mileage grew at about a 0.75% pace in 2017 (source). Our effort to master high-speed rail technology in California was a costly failure–it’s beyond our power.

So enjoy the 50th anniversary of the first moon landing, and start gearing up for the 47th anniversary of the last. Toss a few books on the first to keep yourself warm as you contemplate the godlike power of our ancestors.


Did you mean to post this in the “Book” thread?

1 Like

The notion that scientific advance has slowed down seems pretty dubious.

Yeah, that sounds a lot like the claim that we don’t know how the pyramids were built. Many people interpret that as we couldn’t build the pyramids today when actually we know of several ancient building techniques that were in use at the time. We just can’t tell which one specifically was used and it obviously would be pretty easy to build them with modern techniques and technology.
Could modern engineers build a rocket that flies to the moon? Certainly. Could they build one using obsolete technology? Maybe not. They never learned to use it because why the hell would they if superior methods are available.

1 Like

It’s pretty hard for most of us, maybe any of us, to have a good picture of how things are changing overall. We’re too specialized and have a narrow focus. A big breakthrough in energy or computation or another important area will change perceptions. But even slow incremental progress (that may appear to be small or negative) can lead to sudden unanticipated changes.

Which is not to say that knowledge can’t be lost. I can think of a couple of small examples from personal experience. One being the material that is deposited on the cathode in a vacuum tube to lower the work function. People have little bottles of the stuff squirreled away that was imported from Soviet Russia that works better than what’s produced now. And people don’t know why. (Though could be no one wants to take the trouble to find out. I should also say that was the situation a while ago. Idk what the deal is currently.)

As far as space travel is concerned, Freeman Dyson might be right when he says planets are deathtraps for life. We might never be able to establish permanent outposts.

This sort of OP would be perfect for the forum-linked publishing platform we used to discuss, back when we were still discussing doing those sorts of things.

But “could” is what it looks like to lose a technology. Imagine the day after they finished building the last quality Roman road. Surely the next day they could have started building another high-quality road. They just didn’t, because there were barbarians to fight, and circuses to fund. They could have though. I’d imagine ten years later they could have. But a hundred years later–no. There was a time when they did build Roman roads and a time when they could and then a time when they couldn’t. There wouldn’t have been any way to notice when they crossed over from could to couldn’t though. How long did the Roman engineers and bureaucrats spend telling themselves that they “could” build a new road, but the labor is so expensive, and so many goods move by ship these days anyways? Did they try to build a new road, only to find that even though the project was less ambitious, it took so long and cost so much that they gave up on it? Not because it was impossible, but because there were other priorities. They could do it though, they would tell themselves. Just not today.

This seems a bit like saying education is declining because kids stopped learning cursive. Sending primates to the moon to pick up samples is an inefficient way of exploring space and the reason we stopped doing that kind of thing it isn’t because our scientific know-how is declining but because we got insanely good at computers and robotics. There’s all kinds of incredible space exploration going on right now, but for chauvinistic reasons people don’t get excited when robots collect samples from asteroids or fly past Pluto.

Just in general, using manned space exploration as a proxy for scientific progress seems dubious. I do agree that this Artemis thing seems like a dumb vanity project.

1 Like

True, but it’s still the case that Richard Shelby and Donald Trump still really want to do it and are making sure there are lots of people getting paid to do just that at NASA, and apparently they’re struggling.

What’s the biggest difference between the stuff we currently have no trouble launching (like, the satellites we send to Saturn or whatever) vs. putting people on the moon that causes such a problem? Is it the weight of systems required to keep humans alive the whole time? Is it the extra fuel needed to get back to Earth (from my KSP experience, seems like that shouldn’t be all that much in the grand scheme of things)?

Interesting OP, thanks for posting. I might be back with more thoughts on software after reading that linked article.

Okay, so, regarding this quote, after seeing the author flesh out more of that in the article:

You’ve probably heard this mantra: “Programmer time is more expensive than computer time.” What it means basically is that we’re wasting computers at an unprecedented scale. Would you buy a car if it eats 100 liters per 100 kilometers? How about 1000 liters? With computers, we do that all the time.

Okay, yeah, accurate. My response: so what?

In software, we optimize for things that matter. Usually that means software performing the task it’s supposed to, not squeezing every ounce of performance out of the chips. Would you rather have Microsoft Excel that gives you pretty awesome spreadsheet functionality, or the world’s fastest spreadsheet program that’s limited to 4 cells per file? That’s what I thought. You can have both when Microsoft quintuples the size of their Excel team and charges you $2,000 for a copy instead of $40 or whatever. Is the Excel you actually get so bad? It comes at a reasonable price, most people seem pretty happy with it, and it’s not like it’s sending your electricity bill into the stratosphere. Is that really like a car that’s guzzling 100 liters per 100km (idk silly metric units but that sounds like a lot)?

The guy cries about text editors and how they’re slower than computers were 40 years ago. You know what you couldn’t do 40 years ago? Click a fucking mouse where you wanted the cursor to go. If modern text editors were so unbearable, people wouldn’t use them. They do. You’re just a cranky old codger who yearns for the old days when everyone used Unix, go crawl back to your miserable hole and don’t come out again.

More seriously: when you think about the scale of what huge companies do with software, this sort of problem means millions and millions (billions?) of dollars to them. Think of all the code that Google, and Facebook, and Netflix, and Amazon, and Microsoft, etc etc etc are running in cloud centers around the world and how much it costs to maintain that, how many more CPUs they have to buy and spin up when that software (say, for example, Netflix’s video transcoders that are responsible for an absurd percentage of US internet traffic) isn’t running as efficiently as it could. Those companies are not stupid, they’ve put a lot of thought into how and where their resources are best deployed.

I forget where I read this, and I’m probably bastardizing the original, but I read something about how the average webpage load time hasn’t changed in, like, the entire history of the web - it’s always around, say, 2 seconds (making that # up). The reason being that once you add the newest fancy bells and whistles to your webpage, you certainly don’t want people waiting 5 seconds for it to load and being pissed at the slowness and closing your page, so you work to optimize it, but eventually you get to a point where it’s good enough . Once it gets to the magic 2 seconds, it feels fast enough that you’re like, ok, this won’t bother people, and at this point what’s the marginal return on the time I spend optimizing it down to 1 second? Should I even bother? Probably not.

All the bloat this guy is complaining about is because things are good enough . You can tell they’re not good enough when people stop using a product and Microsoft goes back to the drawing board on whatever’s driving people nuts about the current version of Windows, which seems like it doesn’t really happen anymore? Personally, I feel like Windows 7/8/10 have been some of the best and most reliable operating systems Microsoft has ever made, which seems a lot more important than the shit this guy is crying about like its hard drive footprint . Yeah? Windows 95 was 30MB? Cool, cool, you know what else it did? It crashed all the fucking time.

I feel the need to be a bit contrarian for the purpose of responding here, his points aren’t all bad - but in general, I think he wants to see his programming job like making beautiful art out of code or some shit, instead of being about making products for people. But it’s the other way.

1 Like

The Cold War is the biggest difference, to be quite honest. The US was willing to take pretty outrageous risks to get people on the moon because it was an enormous propaganda victory and also because precision rocketry was a part of national defense. My take here is pretty cynical, but I see the space race as a weapons development program with incredible PR. Bobo’s analogy of the Romans stopping road-building so they could focus on warfare seems backwards to me: today we no longer focus on building giant rockets because delivering an ICBM to Moscow is less of a priority.

Also, NASA is way more safety-conscious after the Challenger and Columbia disasters. A lot of people on both sides of the Cold War died during the space race but it wasn’t on live TV and you could kind of justify that the national prestige of being the first was worth it. Today, who wants to send someone off to risk death just to become the 13th person on the moon? Is there any real scientific rationale, anything that wouldn’t be easier with robots? We should put people on the moon because we haven’t done so in a long time is a bad reason to pursue this.

I think KSP really downplays how fussy astronauts are. The Kerbals just need a drawer of snacks, they don’t need resupplies, nothing ever breaks down, their muscles don’t atrophy in zero-g, you send one out alone in a capsule for twenty years and they don’t mind. We don’t have any idea what happens to people when they’re exposed to radiation outsides the magnetosphere, KSP doesn’t even try to model that.

Just poking in to plug the CNN Apollo documentary. It’s all archival footage, no voice over. Great stuff.

Wharton guys would have it done in a month but it might take longer if their lab assistants were female,young and educationally subnormal.
Too many distractions.

A couple days ago I heard my wife’s aunt, who everyone thought was going to die pretty soon, is cancer free because of some drug that was synthesized just for her. My mom’s alive because heart surgeons did some surgery on her heart going in through a hole in her leg. Early satellites were powered with solar panels that worked the same way as modern ones, but today they are several times as efficient and cost about a thousandth as much.

Necessity is the mother of invention. We don’t need to go to space. Most of the computing going on is inefficient, but most software is used to waste time (especially business software).

Sometimes intentionally. A couple of years after I’d left my aerospace job, I got a call to ask if I knew where the source code for drivers for some old equipment might be. Those LeCroy digitizers were key to what we did, but when we upgraded computers, they wouldn’t work anymore. The first time it happened, I dug out the old 5 1/4 floppy disks and went looking through the Fortran. I found they used software timing loops for delays to wait for the hardware (surprising considering the application). I just found the loops and increased the counter then recompiled until I found values that worked and everything was fine until the next time. As far I knew the disks had been put away for safekeeping, along with copies I thought I’d made. But they’d disappeared.

Put this in the wrong thread…

kerowo Avatar
Jul 21, 2019 17:25:46 GMT -7 kerowo said:
~400,000 engineers at around 20,000 companies worked on the Apollo program. That’s the expertise we lost when the program was shut down; how to manage projects of that size and complexity. It’s doubtful there will ever be a project that size to go to space again, nor will there be enough money to do it fast so we’ll see it half assed and done slow and the apologists will claim there is no need to send women into space when we can send robots. It’s the same half assed, short sightedness that will get us an environmental program that will doom us all to ever more severe weather.

Software is the same problem on a slightly different scale. Faster processors continued to save the field from needing to generate rigorous standards or processes on how to build large scale systems. Because faster hardware made poorly performing code run “good enough” there was rarely ever any economic incentive to build “better” code. Pretty much the only economic incentive in the industry is shipping “working” code first so here we are an entire software methodology built around testing in prod because you can always roll back a buggy release. Have the Mythical Man Month numbers improved much since it came out?

And now listening to 13 Minutes, a podcast about the descent of the Apollo 11 LEM to the moon it sounds like a certain extent of it was making it up as they went and trying to fulfill the goal of a dead president…

Yeah, is it technology we lost or did we lose the ability to run big technical bureaucratic projects? How did we get the Apollo program done in ten years from conception to landing on the moon when it takes over a decade for the SLS to maybe get a launch together? Why is the F-35 at 200% of its budget, a decade behind schedule and delivering a crappy product that is possibly obsolete and over matched before it even gets in the air? When the F-16 project put out bids in 1972 and was in service in 1978. Why does building a nuclear power plant take two decades now when we could go from breaking ground to commissioning the plant in under four years in the sixties?

They really had no idea what the surface would be like and Neil almost beefed it on the landing. I don’t think any manned mission today would have nearly as much risk tolerance.