4 ms·

I would LOVE any thoughts you had on Civboot.org, it sounds like you have a good high level of what roadmap and the general technologies thag would need to be p

I would LOVE any thoughts you had on Civboot.org, it sounds like you have a good high level of what roadmap and the general technologies thag would need to be put in place to make it possible.

kragen 4y ago
I'll take a look! The git repo is a little heavyweight, but it looks like my cellphone connection can handle it. I've spent US$6 on cellphone credit for the last 2 months, and that's about to run out, so I'm getting 60 KiB per second now. It'd sure be nice to have a version of the repo that doesn't include dozens of megabytes of digital audio! I feel like maybe the stated goal, "be able to create a technology stack which can self-replicate with 500 motivated students in 8 years, starting with a US-equivalent 8th grade education," is going to be hard to measure progress against, because every test run takes 8 years. Worse, it takes 500 motivated students with only a US-equivalent 8th-grade education! A much more ambitious goal would be easier to measure progress against. A thing I'm not totally clear on is whether this is an unbootstrapping effort like Primitive Technology and bootstrappable.org, where your 500 motivated 12-year-olds are marooned on a desert island with only the shirts on their backs and the Ark of the Covenant, and have the primary objective of building a jumbo jet from found materials within 8 years, or a bootstrapping effort like Debian reproducible builds. To take an extreme example, if you had a 3-D printer with a "reproduce self" button that needs only rocks, clay, wood, dirt, sand, ore, etc., to print out a new identical 3-D printer, even a single 12-year-old could operate it, even if not very motivated. Would such a "seed factory" qualify as a Civboot? (Properly documented, of course, and with enough additional UI that you could use it to make other things.) What if there's no feasible way to build it by hand from rocks and clay etc. if you don't already have one (or the entirety of civilization)? A Drexlerian nanotech assembler would be such a device, for example. You like Forth. I'm still not convinced about Forth. In the abstract it seems like it should be about as good or bad as C: the base language facilities are slightly more primitive, but Forth permits arbitrary compile-time computation, an advantage C doesn't have. In practice Forth doesn't seem to pay off as a language; I think it's better understood as an operating system for embedded work, with the Forth language taking the role of Tcl or Bash or editor keyboard macros: it's a scriptable user interface, IDE, and "debugger", in the sense that you can inspect the contents of memory. But maybe I just don't have enough experience with Forth yet, and if I just keep practicing, it will become easier and more appealing than C. I've at least gotten over the initial hump of trying to keep all my data on the stack all the time instead of using VARIABLEs (or VALUEs). But I still find writing things in Forth a lot more bug-prone than writing them even in C, much less in something like Python, which gives probably a 2-4 times speedup over C in programming time. And I still find Forth code a lot harder to read than C or assembly. (You might be interested in StoneKnifeForth.) I agree that electronics are probably pretty important, because control is one of the crucial aspects of such a cyclic fabrication system, and one that can dramatically simplify its other aspects. Electronics are an extremely appealing way to do control because electrons are so light, so, at any fixed power level, you can move electrons about 100 times faster than entire atoms. So although there are a huge number of ways to do computation (relays, fluidics, hydraulics, pneumatics, Drexler's rod logic, Merkle's buckling-spring logic, DNA epigenetics, etc.) electronics have a really huge power advantage. (Solid-state electronics especially, though there do exist gigahertz-capable vacuum tubes.) I see electronics fabrication as primarily a question of material processing capabilities: purification, measurement, etching, vacuum pumping, ion bombardment, air filtering for clean benches or clean rooms, stuff like that. Patterning is of course necessary, but the hard part of patterning, at least the way we do it now, is making good photoresists. Is that the kind of feedback you're looking for?
- vitiral 4y ago
  It would, but I don't see it as feasible. Also, I have a thing about incentives, and if the goal is to make it automated then it's probably not going to be understandable. Whereas if the goal is to make it understandable then there is going to be pressure to make it simpler. Simpler technology is also going to have some difficulty with fully automated replication (although, who knows! It may be an important step to that vision too). 500 students and eight years is, I hope, a high end. My true hope is more like 30 and 4 years. Also, the Civboot can be broken down into tasks and built of commodity hardware until replaced, with different groups working on different aspects. If one group can build precision engineering, another group can use equivalent commodity precision engineering to make lenses, ... chemistry, manufacturing, etc can then be bundled together to build a computer chip, that would be a great early success. A full sci-fi school can only happen once all the components are proven and properly documented. In the interim, students can use and learn from built components. Even if this never happens in full, I'm convinced it's pieces would be highly valuable. Edit: also, I can email you the relevant text if you prefer. Most of the size is my podcast and images, which should probably be moved to a separate repo and linked.
- vitiral 4y ago
  Also, I somehow hadn't seen the second half of your comments (was helping a toddler fall back to sleep). I am a fan of what forth can do, but I find the language lacking. With a tiny bit of total code (i.e. 2-3 thousand lines) I've already made a "better" language: github com/civboot/fngi. Definitely WIP, but I'm already writing "code" in it. As for what I'm looking for... anything really. So far the project has mostly been just me, so anything you wanted to contribute would be great. What I thought you would be specifically good at from you comment is outlining the technologies needed to get to precision engineering and semiconductor fabrication, which btw would probably be cold temperature (<300C) thin film lithography on glass or similar. These could include: Creation of vacuums Methods of manufacturing: lathes, CNC, molds, etc and how they have to be changed to do ultra precision manufacturing. Manufacturing and purification of chemicals Manufacturing and testing of optics (for patterning and precision engineering) Precision measurements, including telescopes, oscilloscopes, scales, etc I mean, the sky is the limit. Just understanding the problems and how it is both solved now as well as the history of solving it will be extremely valuable for simplification to the Civboot requirements.

I would LOVE any thoughts you had on Civboot.org, it sounds like you have a good high level of what roadmap and the general technologies thag would need to be put in place to make it possible.

kragen 4y ago

I'll take a look! The git repo is a little heavyweight, but it looks like my cellphone connection can handle it. I've spent US$6 on cellphone credit for the last 2 months, and that's about to run out, so I'm getting 60 KiB per second now. It'd sure be nice to have a version of the repo that doesn't include dozens of megabytes of digital audio! I feel like maybe the stated goal, "be able to create a technology stack which can self-replicate with 500 motivated students in 8 years, starting with a US-equivalent 8th grade education," is going to be hard to measure progress against, because every test run takes 8 years. Worse, it takes 500 motivated students with only a US-equivalent 8th-grade education! A much more ambitious goal would be easier to measure progress against. A thing I'm not totally clear on is whether this is an unbootstrapping effort like Primitive Technology and bootstrappable.org, where your 500 motivated 12-year-olds are marooned on a desert island with only the shirts on their backs and the Ark of the Covenant, and have the primary objective of building a jumbo jet from found materials within 8 years, or a bootstrapping effort like Debian reproducible builds. To take an extreme example, if you had a 3-D printer with a "reproduce self" button that needs only rocks, clay, wood, dirt, sand, ore, etc., to print out a new identical 3-D printer, even a single 12-year-old could operate it, even if not very motivated. Would such a "seed factory" qualify as a Civboot? (Properly documented, of course, and with enough additional UI that you could use it to make other things.) What if there's no feasible way to build it by hand from rocks and clay etc. if you don't already have one (or the entirety of civilization)? A Drexlerian nanotech assembler would be such a device, for example. You like Forth. I'm still not convinced about Forth. In the abstract it seems like it should be about as good or bad as C: the base language facilities are slightly more primitive, but Forth permits arbitrary compile-time computation, an advantage C doesn't have. In practice Forth doesn't seem to pay off as a language; I think it's better understood as an operating system for embedded work, with the Forth language taking the role of Tcl or Bash or editor keyboard macros: it's a scriptable user interface, IDE, and "debugger", in the sense that you can inspect the contents of memory. But maybe I just don't have enough experience with Forth yet, and if I just keep practicing, it will become easier and more appealing than C. I've at least gotten over the initial hump of trying to keep all my data on the stack all the time instead of using VARIABLEs (or VALUEs). But I still find writing things in Forth a lot more bug-prone than writing them even in C, much less in something like Python, which gives probably a 2-4 times speedup over C in programming time. And I still find Forth code a lot harder to read than C or assembly. (You might be interested in StoneKnifeForth.) I agree that electronics are probably pretty important, because control is one of the crucial aspects of such a cyclic fabrication system, and one that can dramatically simplify its other aspects. Electronics are an extremely appealing way to do control because electrons are so light, so, at any fixed power level, you can move electrons about 100 times faster than entire atoms. So although there are a huge number of ways to do computation (relays, fluidics, hydraulics, pneumatics, Drexler's rod logic, Merkle's buckling-spring logic, DNA epigenetics, etc.) electronics have a really huge power advantage. (Solid-state electronics especially, though there do exist gigahertz-capable vacuum tubes.) I see electronics fabrication as primarily a question of material processing capabilities: purification, measurement, etching, vacuum pumping, ion bombardment, air filtering for clean benches or clean rooms, stuff like that. Patterning is of course necessary, but the hard part of patterning, at least the way we do it now, is making good photoresists. Is that the kind of feedback you're looking for?

vitiral 4y ago
It would, but I don't see it as feasible. Also, I have a thing about incentives, and if the goal is to make it automated then it's probably not going to be understandable. Whereas if the goal is to make it understandable then there is going to be pressure to make it simpler. Simpler technology is also going to have some difficulty with fully automated replication (although, who knows! It may be an important step to that vision too). 500 students and eight years is, I hope, a high end. My true hope is more like 30 and 4 years. Also, the Civboot can be broken down into tasks and built of commodity hardware until replaced, with different groups working on different aspects. If one group can build precision engineering, another group can use equivalent commodity precision engineering to make lenses, ... chemistry, manufacturing, etc can then be bundled together to build a computer chip, that would be a great early success. A full sci-fi school can only happen once all the components are proven and properly documented. In the interim, students can use and learn from built components. Even if this never happens in full, I'm convinced it's pieces would be highly valuable. Edit: also, I can email you the relevant text if you prefer. Most of the size is my podcast and images, which should probably be moved to a separate repo and linked.
vitiral 4y ago
Also, I somehow hadn't seen the second half of your comments (was helping a toddler fall back to sleep). I am a fan of what forth can do, but I find the language lacking. With a tiny bit of total code (i.e. 2-3 thousand lines) I've already made a "better" language: github com/civboot/fngi. Definitely WIP, but I'm already writing "code" in it. As for what I'm looking for... anything really. So far the project has mostly been just me, so anything you wanted to contribute would be great. What I thought you would be specifically good at from you comment is outlining the technologies needed to get to precision engineering and semiconductor fabrication, which btw would probably be cold temperature (<300C) thin film lithography on glass or similar. These could include: Creation of vacuums Methods of manufacturing: lathes, CNC, molds, etc and how they have to be changed to do ultra precision manufacturing. Manufacturing and purification of chemicals Manufacturing and testing of optics (for patterning and precision engineering) Precision measurements, including telescopes, oscilloscopes, scales, etc I mean, the sky is the limit. Just understanding the problems and how it is both solved now as well as the history of solving it will be extremely valuable for simplification to the Civboot requirements.