Sam Zeloof has made chips in his garage and posted a whole series about it on Youtube. He bought his silicon wafers, he didn't grow them, and his machines do take up the whole garage - but he did the whole thing himself. Fascinating viewing IMO. I don't know anything about where one would get these garnetty materials you mention, though.

My ergo journey started with similar requirements to yours - specifically including the Y and B keys. Along the way, I learned how important layers are for comfort, ditched QWERTY entirely for Colemak DH, bought a 3D printer, and ended up at 40%. Several years ago, there was a term "1KFH" ("one key from home") people used to describe the amazing amount of comfort they found when they never had to move their fingers more than one key away from home position, nor to move their hands.

I'm not saying you have to change your requirements, now or ever, but I think people who start to make their own ergo keyboards may be subject to this sort of requirements drift, such that if they ever make it to the product phase, their products aren't what they initially expected to be building. And maybe this sort of dynamic is what makes it less likely for the product you are looking for to have been built already.

1. This comment is everything I hoped for but didn't know to expect. Exhaustive, precise, correct, informative, substantive.
2. Yes, dust.
3. MONTY is swamped: 429 Too Many Requests :(

Michael W. Lucas just released a book on how to do this. I've begun but not finished it. https://www.tiltedwindmillpress.com/product/ryoms-ebook/

I have a Folger Tech i3 2020, so named because its frame is made from 20x20mm aluminum extrusion. No bed levelling, no quiet steppers, no all metal hot end, five years old. I've added a part cooling fan whose nozzle I printed off Thingiverse, a janky ring of 24v LED lights, and a cheapo 0.6mm nozzle. Sometimes I have to print part of the first layer a couple of times and move the z end stop to get it right. It takes about 10 minutes every couple of months, so not a big deal.

I say this not to recommend this printer to you today, but to say that even if you don't manage to get perfect prints out of the box, the fiddling it takes to get what you want is probably not that bad.

Besides keyboards, and an occasional toy car or something, I've printed a replacement shade for a little fluorescent kitchen light, an adapter to fit a lampshade that was on sale to a lamp that was on sale, a fancy toilet paper spool, and a custom wrench to try to remove my washing machine's tub. Oh and brackets that hold my cell phone, so I can use my ergo keyboard to type at a terminal on my phone, broadcast my pirate signal, and hack into the Matrix, while riding the bus. :)

Use the Part Design workbench (you probably are already, but no one's said it yet). Sketch a rectangle for the top of the whole tray, not the surface. Pad it down 40mm. Add a draft to set the angle of the sides. Use the thickness tool to dig out the middle of the top face - to a thickness of your exact choosing, which will be consistent everywhere. Now you have a trapezoidal bin.

Then how do you make the separators. Um, draw a sketch with the tee on the inside bottom and pad it... and then the separators don't reach the angled side walls. Oo, how about this: on the inside bottom, draw a sketch of the small square of material at the junction of the tee, and pad this tiny pillar up to the top of the tray. Then start a sketch on a side wall, External Geometry the near sides of the pillar in, and they'll be projected onto the angled side wall. Then loft the two rectangles together. Yeah? Yeah? No. That didn't work. The projection was normal to the angled wall, not to the side of the pillar.

HAHAHA ok. Select a side of the pillar. Pad it, select Up to Face, and pick the angled inside . Presto!

Then stick the lip on top and the grippy bit and that.

I hope this was helpful and entertaining.

I tried Emacs six times before liking it. The time it stuck was when I was editing some code in a language where the include/import statements almost matched the directory structure in the filesystem, but didn't. So, in Vim, I could cursor over an include statement, type gf... and not quite be able to instantly open up the included file. The way gf worked was that it was written in C as part of Vim, and to tell it where to look was a matter of configuration. But I needed a bit of code instead, to make up a couple of places in the filesystem to look on the fly, when I wanted to find-file-at-point (the Emacs term for Vim's gf functionality). Not only was find-file-at-point written in Elisp, but it already had a place where you could hook some custom code in, and documentation about how to do it. The documentation was available inside the editor (as you might expect from using Vim's :help), but it also had a link straight to the Elisp source. I was able to try out my function, change, and try again without restarting Emacs, and debug it step-by-step using edebug.

Anyway - have fun with Neovim. I hear it's spiffy. :)

I agree that you should get a keywell keyboard. I haven't read any specific reviews (I'm down the make-your-own rabbit hole instead), but I've seen some sentiment that the Glove80 is better than Kinesis' offerings, and I believe it's more programmable.

And about that last, if you "have to learn how to type again from scratch," you should use a key layout that will work best for you. This may not be a layout that already exists! Colemak and all its variants, for example, put A and R under your left ring and pinky finger. You might want K and J there instead. Or if it's easy to press the key your left ring and pinky are on, but hard to move them to a different key, you might be OK with A and R. Though Dvorak, for example, has left-handed and right-handed variants, I don't think there are any predefined layouts for people who want to type more letters with their right hand than their left -- or to be more likely to need to move fingers on their right hand more often than their left.

Carpalx is a body of work that lets you define the typing effort for each key, and finds an optimum key layout for you. I haven't used it myself - Colemak DH is a sufficiently high local maximum of goodness for me and I haven't gone down that hill to find a higher maximum yet. But the moment you're in may afford you a unique opportunity.

http://mkweb.bcgsc.ca/carpalx/?typing_effort

If you do end up making one, or having one made, you might be interested in something like the Concertina.

I've been using a tshort dactyl manuform 4x6 for 5 years now, having never planned to use it for even 1 year. I only commonly use the two innermost thumb keys; I didn't think I would like the thumb cluster from watching a video of someone typing on it, and I indeed don't like the thumb cluster. The switches are Kailh Brown; one of them started to stutter and I replaced it with ... a TTC Brown or some such.

I've printed, but not finished, a Splaytyl. I think it's going to feel nice, but it's only 4x5, and I'm nervous about not having Tab and Enter on the base layer.

I would pull on that thread. That is, in your shoes

Directions unclear; shoelaces tangled

Secure Scuttlebutt is (was?) a protocol for high-latency communication between occasionally-networked humans. Pro: https://scuttlebutt.nz/; con (not read in detail): https://derctuo.github.io/notes/secure-scuttlebutt.html. I think it was supposed to be able to spread messages over Bluetooth, assuming a sufficiently connected web of nodes between person A and person B. Public keys were identities, and were bound to devices; unfortunately people may have multiple devices, or change devices over time, so this was a hindrance.

IPFS was supposed to be the Interplanetary File System. I think that was just because whatever pieces of content you ask for, you also cache, as part of the design: you keep a copy on the near side of the small high-latency pipe. But that's mostly about file transfer, not interactivity.

UUCP was definitely made in a time where a latency of days for delivery of email or netnews was common.

In the early days of CGI, the Web was just one way people imagined interacting with applications; another way was email. RFC 3834 has some recommendations for people who are going to automate email responses. There used to be services you could email a URL to, and receive the web page back as an email.

Using ed (in my experience) involves looking up the screen, or up the roll of paper on your teletype, to see what the lines of your file were, and imagine what they are now, given the changes you've wrought to them since they were printed, and then turn them into what they should be. With Mars rovers you have a simulation that you issue your command to, before sending it off to Mars. With correspondence chess you might keep a physical chessboard for each game you have going, and/or send a form back and forth that keeps track of several moves.

People used to do computation at universities and businesses by writing programs at their desks, submitting them to be typed on punchcards, and receiving printouts some time later. They would "desk check" their programs before sending them in, because each compute job took a couple days to come back.

I mention all these because, in an extreme censorship environment, any local state (session history on paper, an app on a smartphone, an odd device) might not be good to have around. So usability may require reducing the total amount of state that a command carries. The current working directory at the time a command is run changes the meaning and outcome of the command; you may not remember that directory in a day or two. The vocabulary and syntax of command-line switches are easy to look up in online manuals - but are there offline manuals? I don't know if this avenue of inquiry helps you, but it's interesting to think about for a moment.

No, but I've been typing on one for five years anyway.

The concave keywell shape of the Dactyl Manuform is more ergonomic for me than flat keyboards, even when they are tented with an opening angle. (I had a two-piece flat keyboard, the Keebio Fourier, before this Dactyl Manuform.) I expect it to be more ergonomic for anyone.

The thumb area of the Dactyl Manuform is constructed with the idea that you will actuate the keys there using the pad of your thumb, not the side. The original Dactyl and the Kinesis Advantage 2 have a thumb cluster which is raised up relative to that of the Dactyl Manuform, and I think this is more correct.

With your HHKB and Realforce collection, you might be interested in the Dometyl, a keywell keyboard made to support capacitive rubber domes. Its three thumb keys are probably angled better; but I have not yet finished building my Splaytyl (another variant in the same repository).

i got paul through the flames! but he dies if you feed him too much

It took me a couple months to stop thinking about layers.

My first move toward ergonomics was a Keebio Fourier 40%. It had been a few months before that when I started using Colemak on my laptop keyboard; I did that using Windows keyboard settings, and I was taking notes in meetings, so whenever I couldn't keep up, switching back to QWERTY was a hotkey away.

After switching to the Fourier, and iterating many times on the QMK settings, that was the month or two where I had to think through all the control keys, symbols, and function keys I was typing. I didn't quit typing on other keyboards, although I typed on my Fourier as much as possible; and I have not ended up forgetting anything I learned before.

I'm now on the precipice of moving from a 4x6 Dactyl Manuform to a 4x5 Splaytyl (if I can find those dang parts and get the thing built!) and that's too small to have things like Tab, Enter, or the backslash on the home layer. I'm nervous. I've tried making a 4x5 layout for the DM, but haven't ended up sticking with it. I couldn't really get used to keys that are physically there, not doing anything.

The Glove80 is keywell-shaped, and is a product. https://bastardkb.com/ has fine Manuform-thumb-cluster keyboards.

I have printed a Dactyl Manuform; its thumb cluster has six keys, but is quite different from the Kinesis, I'm afraid. I think the original Dactyl was an attempt at making a more portable Kinesis Advantage. I guess you can maybe generate a Dactyl design at https://ryanis.cool/dactyl/ ?

I'm afraid my advice is a few years old because I haven't kept up. The Dactyl Manuform was supposed to be a stopgap, but it's worked well enough that my focus has drifted to other projects.

See also kbd.news if you haven't before. I used to think, why is this dude making his own website when everybody's already on r/mk? ... And then Reddit sold out.

Colemak DH represent!

That's beautiful!

I have heard that there is some arrangement whereby you run FreeBSD on your hardware, with Linux in a bhyve virtual machine; you hand your physical wireless card into the VM, where LInux's drivers can talk with it; and route packets from the main system through it. -- Ah, https://blog.desdelinux.net/en/wifibox-the-project-that-allows-you-to-use-linux-wifi-drivers-in-freebsd/

To try to set up such a thing as your first entry into a BSD might be frustrating, but if you kept on until it worked, you would definitely have dived deep. (Dove? Doven? Diven?)

I'm a long-time Linux user now using FreeBSD on my home server. The first few times I looked at BSDs, they seemed old and stale, like nothing was happening there, and the coreutils were less comfortable to use because they were missing some switches. But what I've learned is that FreeBSD builds incrementally, without undermining itself, and my own understanding of it can do the same. What I've learned about previous versions of FreeBSD is more likely to still be true about the next version of FreeBSD.

BSD people often mention how the BSD in question is built as a whole, not cobbled together as a distribution. This difference can be stated far more quickly than it can be fully understood: like a culture, of which you gather a nuanced understanding from a broad survey of its literature, rather than a movement, whose goals are painted in broad strokes by a manifesto.

Anyway, welcome! My experience has been that #freebsd on libera.chat is more lively during US daytime hours than later at night. The Handbook is definitely your first documentation stop. ZFS, with its snapshots and replication, seems to be the most-hailed feature of FreeBSD; DTrace didn't even make the top 10, but when I didn't understand why NFSv4+Kerberos was failing, it was indispensable. Have fun!

Uh, I think the Glove80 uses Choc switches, right? For heavy tactile in Choc you would want Burnt Orange. Not sure whether that's an option they provide or what.

Bastard Keyboards -- I've talked with Quentin and he seems like a cool guy. He's an innovator in the use of printed-circuit boards for keywell keyboards. That's important because it makes keywell keyboards much easier and quicker to make, without the huge cost associated with polyimide flexible PCBs. He has high quality standards, too, in my limited experience of his products.

haven't tried it but i think the thing i've read about before is https://archivebox.io/