Shell World

Megastructure for sale! Get yer sci-fi megastructure here! Any authors that want to play with this idea welcome.

This is a write-up of an interesting idea I had back in August 2020. It’s for a particular type of sci-fi megastructure which should be stable (at least to back-of-the-envelope calculations) and would make an interesting setting/location/super-architecture-style for any authors who are interested. Not practical to build anytime soon, but interesting.
I’m not sure if this is original or not (I haven’t looked very hard.) If it is, I’m staking my claim! Publish or perish!

Shellworld Writeup

The basic idea, as described in the document, is that for a thick enough, large enough shell of matter, the force of gravity can balance internal pressure. The radius divides out, so this structure can be built without tensile stress to arbitrarily large radii, allowing the construction of massive amounts of habitable volume.

Interesting Reading

Some interesting things I’ve been reading this year:

The Wright Bros by David McCollough is an interesting biography of the inventors of controlled heavier-than-air flight. One fascinating part was a reference to a speech made by Wilbur Wright at the Western Society of Engineers.

Link to my copy of the speech: “Some Aeronautical Experiments”, Wilbur Wright, 1901:

Link to where I obtained it from: Link
Link to the biography: The Wright Brothers, David McCollough

This is one example of what a first rate mind approaching a problem looks like. Wilbur makes it clear that the principal difficulty that he and his brother were tackling at Kitty-Hawk was the problem of controlling the vehicle while in the air. They built up the knowledge to do so with an extensive series of glider experiments. The experiment-heavy, experience heavy and iterative nature of their self-directed program contrasts with other efforts to build flying machines at the time: A large upfront investment made in an expensive machine which promptly crashes.

While the Wright Bros faced some pretty stiff incredulity and cynicism, most of that cynicism came from the world of the press and humanities. There was a fairly serious and supportive community of engineers who were making efforts to tackle the problem of heavier-than-air flight. (The Dayton Daily news refused to acknowledge their accomplishments, even as they were doing laps around Huffman Prarie. It fell to a Kentucky almanac run by a beekeeper to provide their first publicity!)

Book Workbench Python Scripts

Link to the scripts

Book Dumping and Book Binding Scripts

Aaron M. Schinder

18 Mar 2021


These scripts are intended to dump image files from pdfs and bind image files into pdfs. I wrote these primarily to convert pdfs from the highly compressed format they use to less compressed but more readable pdfs built from the images.

The various Linux pdf readers have difficulty unpacking books. It’s a known problem. (It takes a very long time to load pages in whatever format or scheme uses internally)

The idea here is that hard drive space is cheaper than time, especially the loading time for the pages which may make paging and skimming through a book untenable. Paging through images can be done almost as fast as flipping through a physical book, so it doesn’t impede using them in a similar manner.


These tools were written to work on Linux (I don’t know if they’ll choke on Windows or not – may require some fiddling.)

They require the imagemagick command line tools (convert). Imagemagick may need the settings adjusted in /etc to avoid running out of memory.

They require pdfinfo and pdfunite and pdfinfo utilities from the poppler package/library.

Contents : Dumps a pdf to image files in a directory named after the pdf file. : Iterates over all pdf files in the directory it is called in. : takes all the image files in a directory and creates pdf fragments in stages. The pdfunite is called to concatenate all the pdfs into a new bound pdf. (Warning, it will name the pdf the same as the directory name: shuffle the files around to avoid an overwrite.) : takes all the book image directories and makes pdfs from them.

February 2021 Machine Shop Adventures

I recently bought some stepper motors from Sparkfun. I have two others floating around somewhere – they are of the NEMA-17 type (which specifies the face and shaft dimensions and mounting holes.) Mine are m-3 instead of 4-40 mounting holes, but otherwise standard.

I’ve often longed to do something with these motors – to build tools which can build better tools. My initial attempts to build motion stages involved buying stock parts from McMaster and attempting to 3d print all the connecting parts with a Makerbot Replicator 2X. Unfortunately the 3d printer has only about 0.04 inch accuracy, which isn’t enough to make motion stages that won’t bind, even when just aligning steel rods.

In my quest to become able to build serious tools, I’ve acquired a machine shop. I’m still not sure quite how to “climb the tech tree” from poor tools to better ones (more accurate, more precise). How *did* we get from hand tools to the bridgeport? How do you make something more accurate with something less accurate?

(edit, in response to the comment below: Since trying to build my first stages, I have done a little research on some of the steps in this process. I think one of the first ones involved Maudsley in England creating a lead screw using a self-propelled inclined knife on a flat slide climbing against a brass-rod workpiece. Also the Whitworth grinding proceess (which I’ve used myself on 3 pieces of steel – incredibly tedious) to get flat planes. That gave him a master leadscrew which eventually gave him a lathe, which gave some degree of positioning accuracy. I’m not surprised to see interferometry in there. Still haven’t read up on how they managed squareness. I’ll have to watch more of Dan Gelbart’s videos, since he’s been through the process I’ll eventually want to go through.)

Anyway, step 1 would be to build some motion stages so that I can build a CNC thing. (I’m thinking small, high-RPM/low tool force carbide bit CNC mill or something…)

Pencil scribbles

I am not my grandfather. My grandfather could do beautifully fine hand-drawings with drafting tools. I’ve inherited this old map of his alma-mater, Tri-State College done by him during a course on engineering drawing (will include eventually). Thomas French, Engineering Drawing – it was tucked into his textbook.

Hand drawings are useful: They’re faster than CAD when you’re still trying to figure things out, and may be entirely sufficient to plan out a part. Here is that hand-drawing cleaned up and colorized for visual interest.

The big idea

Anyway, this was the plan. I ordered a bunch of stock parts from McMaster, I made one attempt last week to cut the dovetail slide from a piece of aluminum. I figured, aluminum is softer. I can practice on it without blowing up my tools. That worked until I attempted to use the dovetail cutter. Advancing about 50 thousandths per step (way too fast), without using adequate lubricant (Dad suggests WD-40 when cutting aluminum. *Not* when cutting steel – then it would carbonize and work-harden the part.), I ended up blowing up the dovetail cutter in my face.

Aluminum doorstop that also handily stabs you

I decided, since getting the part to that state took 5 hours, to bite the bullet and just do it in steel this time. 5 hours of end milling and a nautical ton (it’s an English unit, honest!) of chips later, I had the slide squared and the square-cuts done with nice surface finish. Can’t feel any roughness with my fingernail (the fingernail test).

Beautifully squared piece of steel, ready for wrecking

I painted some Dykem blue layout fluid on the corner in an attempt to see my cuts.

The wax goop on the tool was difficult to pry off. Tools are often dipped in this waxy crud to prevent corrosion. I managed to get it off by submerging the tool in a pot of boiling water – this worked and allowed me to easily remove the wax, which is important when attempting to get the tool into a precise starting position.

This shows the tool ready to cut. This time I was paranoid, listened very carefully to the horrible noises the cutter made as I advanced, and probably saved myself from a blowup at the very end. I advanced about 5 thousandths per step towards the end, and only about 0.2200 of my planned 0.2500. On the last cut the tool made a horrible racket on the last inch of the cut. On attempting to start the back-side, the tool wouldn’t cut at all. It had dulled itself fairly quickly.

It was a high-speed steel (as opposed to the cobalt steel) 45 degree, 1 3/8″ diameter dovetail cutter from McMaster Carr. (Not sure if I recommend them – they dull too fast.) I’m waiting on a new cutter to finish the job, and some carbide inserts for a 60 degree insert dovetail cutter from Dorian Tool by way of MSC.

One random trick that I had heard about watching youtube videos of people doing far more intricate things: Apparently when cutting steel, you can pay attention to the color of the chips. Straw-yellow means that the chips are getting somewhat hot as you are cutting. If they start turning blue, it means you are work-hardening the part and generating too much heat, and need to proceed (with more coolant, less agressively, at lower speeds).

South Bend Lathe Model SBL400 Parts List

I recently purchased a used metal lathe: A South Bend model SBL-400, from Jones Machinery in Cinncinatti.

Unfortunately, I did not have an opportunity to inspect the lathe very closely until now, and there are too many things wrong with it to make much use of it. It’s been used hard over the years, and the cross-slide skips. There’s some sort of collet holder wedged into the spindle and I’m not sure how to remove it. The foot-break won’t engage. Etc.

It’s a good model, but this one would be dangerous to attempt to operate, and I’d have to sink a lot of money into repair.

As part of the purchase, I bought a parts list that I couldn’t find anywhere online in digital form. Here is the parts list for anyone else out there lucky enough to have an SBL-400 in better working order:

SBL Parts List

Comment on James Cambias’s Blog

This isn’t entirely aligned with the topic of his post, but I saw something in the news about another SETI search coming up empty. I don’t think this indicates much about the absence of alien life in the universe, for the reasons in my comment:

One of the articles mentioning the seti survey, but not the original:

(PS: James Cambias is the author of several science fiction novels that I’ve greatly enjoyed. A Darkling Sea is highly recommended. (One part great novel, one part novel-length panning of Star Trek’s prime directive.) Arkad’s World also makes some interesting points between the lines.)

My Post:

This doesn’t really address your post directly. I saw that article that floated recently about another SETI survey of some small fraction of the sky turning up empty again.

It seems popular to assume that if our radio waves could have traveled 100LY in principle, that we’re visible out to 100 LY, and that we could see aliens broadcasting like we do from such a distance. I think interstellar communication is more difficult, and necessarily more directional than most people assume.

The Galileo probe was an interplanetary spacecraft that operated around Jupiter. Galileo’s high gain antenna failed to deploy. Earth was trying to find the 20W omnidirectional low-gain antenna to talk to. It’s amazing that they could do it at all.

I was doing some physics doodling to get some numbers back into my head: I recall we managed to laboriously drag a 100 bit-per-second signal out of the noise floor from the Galileo probe at Jupiter. It was an S-band transmission, 20W, which amounts to 2,6E-24 W/m2 on average here at Earth. We had to coordinate several dishes with detectors cooled to 11K to obtain that signal, and we could only do it because we already knew Galileo was there and where to point the dishes!

Those big antenna dishes on planetary probes are not there for cosmetic reasons!

A 1GW omni-directional transmitter does roughly 100x worse at 5 LY, the approximate distance of our nearest-neighbor star. Aliens right next door couldn’t pick up a massive omnidirectional transmission, nor could we. Practical interstellar communication *must* be directional.

With a 1km dish in the X-band (8-12GHz) we can get gains of something like 2E9, by creating a beam that has a half-angle of 4.57E-5 radians. A 1GW sender could then be heard at “Galileo levels” at 25000 ly, or 1/4 distance across the galaxy. A 1kW sender can be heard at 25 LY at those extreme-limit levels, but the beam would only be about twice as wide as the orbit of pluto at that distance. For economical transmission powers, ranges are in the local-stellar neighborhood and necessarily pointed at specific stars. No one else could intercept that beam unless along a direct line of sight.

Your optics don’t have to be as extravagant in optical-laser wavelength ranges. An equivalent optic to that 1km dish for a 535nm laser is a 1.5cm lens: easily doable. You’d have to have a power of >800W to outshine the sun at a 1nm bandwidth, also doable with pulse lasers.

Anyway, I suppose what I’m saying is that it isn’t too surprising that a radio survey is coming up empty: Doesn’t indicate anything even if the galaxy is full of civilizations, other than that we’re not on the “point to” list. (The massive time disparity between “has radios and lasers” and “is moderately intelligent” that you discuss here seems more important.)

Comment on Isaac Arthur Video

On Isaac Arthurs thought provoking video on Consciousness and Identity:

My comment:

I dunno that you can place an upper bound on number of unique personalities. Configuration space blows up fast: For certain definitions, you must end up with more possibilities than raw materials. (Number of ways to arrange N balls is N!, which is > N when N>2). You can store N bits on your hard drive, but there are 2^N possible hard drive states.

Under one physics definition of identity, all electrons are the same. Drop the resolution, and you can say one bacterium is pretty much like all others sharing the DNA. A lot of lower animals seem like they approach life in much the same way and have more or less commensurate experiences. (Sphexic – insect like. All bees of a certain species will execute the same orbit when identifying something they want to land on.) The more complex the mind, the more space there is for these minds to be different.

Going to start posting some of this stuff here: I comment a lot on other people’s blogs, but my own blog ends up being pretty bare. If I spend all my time talking elsewhere, I won’t have a lot of content that accumulates here.

How to Internet 101: Backups

Rule 1 about anything that touches the internet: This goes double for servers that are on all the time and configured by amateurs (you and me): KEEP BACKUPS.
Rule 2 of the internet: KEEP BACKUPS!

Always back up your data, your server configuration, and your databases. That way, when you view your site one day and discover it’s been overwritten by SEO spam selling viagra in chinese, you’ll be able to wipe it and restore.

If you’re handling customer data or people’s credit cards, you need a higher standard of security. If you are only serving content though, the easiest and best security is just keeping backups. You can always nuke it all and start over. With Linux, nuking it all and starting over is an hour exercise that doesn’t even require calling the Microsoft mothership for authentication.

backup the /etc directory (config files)
backup the /var directory (where you’ve probably put your mail and website files)
backup your /home directories

Some helpful database backup stuff for mysql:

Dumping a database to a backup file

msqldump -u username -p database_name > file.sql

Restoring a dumped database file to the database (first create the database)

mysql -u username -p database_name < file.sql is getting censor-happy again

My home-bases on the internet are hosted on either server computers in my house, or servers that I rent from vanilla-hosting providers. I recommend Linode for hosting, and Namecheap for domain name registration. So far, they haven’t burned me.

Unfortunately, our cultural civil war has unleashed a bunch of censorious hyenas on our companies, and the interent as a whole. Payment processors, blog-service providers, hosting providers, and other services are being systematically denied to people because of their politics, or just the information they provide and the people they “associate with”. This is possible, because people have placed their content into the hands of a few centralized service providers.

Running your own server has never been more important, to preserve the freedom of the internet, to keep yourself from being silenced, and to keep your content from being arbitrarily stolen or destroyed by the asshole companies you entrusted it to.

Running your own blog on your own server is not difficult. If the email tutorial below scared you, don’t let it: most things aren’t as mind-numbingly complicated as a Postfix/Dovecot server. Your own instance of the wordpress blog software can be run on a server that you control. There are also many other php applications out there that are freely available to use (wikis, other blog software, php forums, etc). The whole infrastructure of the early-2000s internet is mostly available for free – all you have to do is sacrifice some SAN points to the penguin gods. But I think it’s well worth it in the end to have your own home online.

A comment I posted back in 2018:
On Adaptive-Curmudgeon’s blog (very fun blog, check it out): Link

Ditto. I moved my blog to a server that I rent. (I keep backups and can move it again to another server if necessary). WordPress’s recent misbehaviour though leads me to wonder if I should back up all my posts and media onto a different format entirely, in case they start trying to go after independent instances of their software through some backdoor.

I was also recently IP banned by google. That was a fun experience. I was looking up academic papers, but I must have looked up the wrong one in the wrong way (spin up the tinfoil beanie: there must be some deep dark secret about superconductivity!). It turns out there is absolutely no customer service number to call, and absolutely no one to appeal to. Not even an automated robot, as infurating as those are to deal with. If you’re banned by google, you’re banned. I called up my internet service provider to get them to reset my IP address, and they treated it like a federal crime to try to evade a google-ban. Something must be deeply wrong with me, my computer, or both, if I happen to be *disliked* by a mindless robot. Send the FBI! (I called sales the next day, and salesmen seem to have a more pragmatic attitude towards what they owe their *paying customer*, especially since I had one toe out the door at that point.)

While it is *thankfully* possible to go through different services, it’s nevertheless unsettling when the owners of a near-monopoly act like assholes. Just because it is possible to defend against assholish behavior, doesn’t mean that the world isn’t a less friendly place because of it. It could be one definition of a golden age: For a few brief decades, people cut each other a tiny break, do what they promise, and stop acting like jerks! Civilisation flowers.

I agree, learning how to defend against it and owning your own platforms are vitally important. It’s somewhat depressing that so many people have followed a convenience gradient down into placing all their content in just a few oligopoly platforms that they don’t own or control.

Email Server Instructions

I’ve written up a very rough draft of instructions on how to set up your own email server. These are instructions for setting up a Postfix/Dovecot server with OpenDKIM authentication on Ubuntu and Raspbian.

Link to Tutorial