Hacker News Reader

Comment by entee | Hacker News Reader

I often wonder as I am building/fixing something (often my bicycle as it happens) and I happen to toss an extra screw in the trash how miraculous a simple screw of this type would have been to a middle-ages craftsman. The precision of keeping all those threads perfectly spaced is one thing, but the technology to make them in such volume at these sizes (often quite tiny), and in volumes so as to make them essentially disposable is a true marvel.

We think a computer is an impressive piece of technology, but many of the things around us that we consider less than mundane are just as incredible if we look just a couple hundred years back.

View on HN

"The precision of keeping all those threads perfectly spaced is one thing"

Making the first precision screw requires a "screw originating machine".[1][2] This is a special-purpose device which can make a better screw than the ones used in building it. Then you can install the first precision screw in a lathe and make more screws in harder metals. This is how precision is bootstrapped.

[1] https://chestofbooks.com/home-improvement/workshop/Turning-M... [2] http://collection.sciencemuseum.org.uk/objects/co46546/tool-...

> "The precision of keeping all those threads perfectly spaced is one thing"

> Making the first precision screw requires a "screw originating machine".[1][2] This is a special-purpose device which can make a better screw than the ones used in building it.

I hadn't realized screws would be needed to create a screw-making machine. I love that we have machines that make better versions of their own components, something I never even thought to think of!

I don't have a source, and maybe I'm misremembering, but I think compilers (for software) and CPU production tooling (for hardware) both bootstrap off their own performance.

A better compiler can compile itself more efficiently, same for CPU/chip layout systems.

You’re exactly right about this. Of course compilers which compile themselves come with their own risks! https://dl.acm.org/citation.cfm?id=358210

Risks which can be ameliorated: https://dwheeler.com/trusting-trust

The original version of the C# compiler was written in C and C++, but today's version is indeed itself written in C#!

https://stackoverflow.com/questions/27513234/in-which-langua...

Often, the first bootstrapped is the bare minimum to get basic functionality available, to then build upon what features you can support.

Its been a while, and I think it was gcc 7.x timeframe, and IIRC, you had to have at a min, gcc 4.x available to bootstrap, and it took something like 3 separate bootstrap iterations to get the final compilers was like 4-6 hours if memory serves.

> compilers

Absolutely. Otherwise, all compilers would be written in Assmebly. Most compilers used today are written in C and the compiler for C (GCC) is written in C itself.

The C language was also written by Ken Thomson, because he was unhappy with the BCPL language, so the compiler for C was originally written in B. So on and So forth.

More Recently, the initial compiler for Rust was written in OCaml but is now written in Rust itself.

Slight nit: GCC has been written in C++ (not C) for several years now.

This is accurate, but compiler runtime efficiency and compiler effectiveness at producing efficient programs are inherently orthogonal.

There is certainly a major milestone in each new language though, when it becomes feature complete enough to write its own compiler.

Amusingly, most commercial screws today are made with rolled threads, where the wire is rolled between two ridged dies. But making those dies is a pretty high technology in its own right. The use of screw machines is now limited to making specialized or super-precision parts.

This is kinda inherent in everything, our tech generally moves upwards based on iterations of improvements. Even cheap mass production shitty stuff is still an improvement. Or do you mean exactly screws?

Really when you get down to it, like computers, tools are used like 30+% of the time to build other better tools. Jigs, more accurate cutters, measurement devices, etc. With code that's things like test frameworks, rpc mechanisms, version control, code gen, etc. In fact in software engineering you're usually seen as a better engineer if you build things indirectly (frameworks) instead of mass production of the solution.

Check out the home machine shop series to see how you'd build up a tech stack from basically scrap metal and wood and Sand.

The Metal Lathe (Build Your Own Metal Working Shop From Scrap Series Book 2) https://www.amazon.com/dp/B007USU8HU/ref=cm_sw_r_cp_apa_i_Gh...

The lathe is pretty much the ur-machine, and a lathe is really just screws

There is an old Jewish tradition (written down in Pirkei Avot ~1700 years ago) which lists the objects that God created "on the eve of the Sabbath at twilight" after finishing the rest of creation. Most of them are religious or mystical in meaning, but the final object listed is "[the original] tongs, for tongs are made with tongs."

here's a link to the text https://www.sefaria.org/Pirkei_Avot.5.6?lang=bi&with=all&lan...

That's really interesting. The Tanakh never mentions an "evening and morning" for the Sabbath, unlike the first 6 days.

Good catch. That's because we're still on the 7th day according to the Tanakh.

The form of the verb translated that God "rested" is in fact progressive in the original Hebrew.

https://wol.jw.org/en/wol/d/r1/lp-e/2011525

I spent many years building a database backend from scratch. Right from the beginning, you need a fast way to look up a file address given a simple key, and that mechanism needs to be persistent. Thus to build a database you first need ... a database. Bootstrapping this thing was tricky.

The Technium is a superorganism.

Which is incidentally yet another reason why it's very important to ensure stability of our technological civilization. We've pulled ourselves by the straps of our boots a couple hundred meters into air now; it's not safe to stop pulling. Given big enough shock to economies and populations, we'll regress not a decade or three, but all the way to the beginning of industrial revolution - you can't build the technology we have now from scratch, as it's full of cyclical dependencies. It would need to be rebooted from scratch.

This is one of the most fascinating things I've read on this site. Do you know of any resources that go into this in more detail?

See resources around the thread about precision manufacturing. Consider how Industrial Revolution started thanks to easy access to high-density energy sources like coal, and how we've mined and burned pretty much everything that's accessible by hand. Observe how modern mining requires huge amounts of energy and sophisticated technology. It's all intertwined.

Consider also the amount of people needed all across the industries and supply chains of any product you know. I don't have hard sources, but I fondly remember this essay by Charles Stross[0]. TL;DR: how many people does it take to maintain (not improve) current technology level of our civilization? Charlie puts it at 100 million to 1 billion.

And now think of the economies and infrastructure needed to just feed these people. We hit 1 billion around 1804[1], which is far in the industrialization process, and most of these people weren't working to support the technology levels anyway.

For some insights, I recommend tracking down and watching Connections[2]. It's an old show, from the era where TV shows actually made sense, and it drives home just how much our current technology is dependent on right combinations of social, economical and technological conditions.

EDIT: fixed site reference in [2].

[0] - http://www.antipope.org/charlie/blog-static/2010/07/insuffic...

[1] - https://en.wikipedia.org/wiki/World_population_milestones

[2] - https://en.wikipedia.org/wiki/Connections_(TV_series); I hear you can find it on a site that starts with "daily", and ends with "motion.com".

Thanks! I'll check those out. It would be amazing if someone had put together a piece that tied this all together.

In some ways this is what has happened with the transfer of manufacturing to China. Building things requires building components, and their components, and their components, and so on down. That whole stack is withering in western democracies, so it’s incredibly difficult to “bring back” something like iPhone manufacturing without retaining a deep dependency on China anyway.

This is the answer to the question "Why not build more Saturn Vs". We no longer have the tools to build the tools to build the machines to build the parts of the Saturn V.

Why not?

Because "unused muscle atrophy" applies very much to our technology as well. A lot of knowledge surrounding technological processes isn't really captured beyond institutional culture and individual memory. So if there isn't demand for some specific technology, the capacity and knowledge to build it will wither and die with the organizations and people who built it.

We can't do a Saturn V today. We might be able to do a Saturn V equivalent, though, as both NASA and SpaceX are retracing the steps towards heavy lifters.

You should write all this up on a single place, it's very interesting.

I just might put it up on my blog. Need to collect references for the things I've been saying first.

I first need to finish and ask someone to proofread my long overdue post on my views about advertising industry.

I can proofread if you want.

Thanks. I'll send you a draft later tonight.

http://jacek.zlydach.pl, yes? I eagerly await this piece. I can think of a lot of people (and groups) who would find it enlightening!

Yes. I see you have an e-mail address in your profile; I'll drop you an e-mail when I'm done with it.

Thanks. Already subscribed to your feed :)

Hey person, I hope it wouldn't be to presumptuous for me to also ask for an email about this? I'd be very interested in what all you have to say.

Not at all! But since you don't have your e-mail published in your profile, you need to mail me so I know where to send the notification :).

I remember reading that some of the textile techonology that enabled the arctic/anarctic expeditions in the early 20th century has already been lost.

They're having to analyse the garments in medical scanners to work out how the fabrics work.

The idea that technology that enabled the equivalent of the space race could be lost within a century is quite sobering.

This. And if one still struggles to understand the importance of institutional culture + individual memory, I would like into invite you to the following thought experiment.

“Imagine waking up the next morning and the world totally forgot all knowledge on Linux Kernel, and you are now tasked to release the next major version with major changes / improvements”

And remember the documentation and tests we have on Saturn V is certainly not as great as what we have on Linux kernel.

Perhaps more critically, we've burned most of the easily-accessible fossil fuels, which means that if we drop below a critical threshold we won't be able to reboot from scratch because there will be insufficient high-energy-density material available.

That too. I'm increasingly strongly of the belief that ensuring the stability of what we have now is the most important thing. We can keep fixing things incrementally, but if we break our civilization now, we condemn the next couple dozen or more generations to medieval life on a severely damaged planet.

That's why it pains me when I see activists crying for correcting the wrongs at all costs, "justice" at the cost of "order". It's ultimately a self-destructive approach that's akin to cutting off the branch you're sitting on, after first setting it on fire.

The biggest threats to the current order today are not the people calling for more "justice", or more "social justice", however.

The biggest threats to the current order are companies and politicians that refuse to address our unsustainable level of pollution, which is pushing our climate out of balance.

They're the people advocating for a "starve the beast" approach to governance, literally trying to create a financial crisis just to get a few percent taxes shaved off their bottom line (and probably think they can profit off the crisis too).

They're the people pushing for increased gerrymanding and census manipulation, which radicalizes politics on both ends of the political spectrum, shuts out people in the middle, and erodes faith in our institutions on the long term.

These things are by far the biggest threat to the long term stability of the current order in the US.

I agree. And I didn't mean to say that pursuit of justice is a problem, but I do have an issue with calls for justice at the expense of everything else that I sometimes hear from the Internet pitchfork mobs, and I worry that it may gain momentum at some point.

I wonder how difficult it would be leapfrog oil and go straight to nuclear power.

If any necessary and non-substitutable material or equipment is based on petroleum chemistry, then we won't be able to do that. I'd love some nuclear engineering experts to chime in on it here; my uneducated guess is that it's extremely unlikely that there isn't a hard petroleum dependency somewhere in there.

Just think of all the protective equipment. What do you think it’s made of?

I think it would be a lot easier to go with internal combustion of ethanol and biodiesel (and possibly wood gas) first.

The coal-powered steam locomotive era and the nuclear power era overlap. We almost did go from coal to nuclear. Especially in the UK, which had little domestic oil or gas in the 1950s.

This is also why the protection of supply chains and international trade is essential. The cause of the Dark Ages was the collapse of the security surrounding the Roman road network, making trade hazardous, and subsequent reversion to local production of everything resulting in the loss of technological prowess.

In our current economy, that reversion would be even more painful. In the short term, as much as I hate the US military, their "freedom of navigation" guarantees are essential. In the long term, every nation/bloc should make sure it has the materials to re-bootstrap if necessary.

Here in Korea, we just had to scramble because Japan decided to limit photoresist and high-purity HF exports. I count the scramble as a good thing for national self-sufficiency.

My apologies for going way off topic.

Your first paragraph is very interesting. Got a link or book you could recommend to learn more?

The Hardcore History episode "Thor's Angels" goes quite a bit into that territory. It covers period between the fall of the Western Roman Empire until the formation of the Holy Roman Empire.

For fans of that podcast, it's one of the best single episodes. Dan Carlin has a particular style, which is not for everyone, but fans know what to expect.

https://www.dancarlin.com/product/hardcore-history-41-thors-...

There’s a great book about exactly this - including an attempt to help you reboot it by giving people the Igor clues in the right order.

The knowledge, How to rebuild the world from scratch https://www.amazon.co.uk/Knowledge-How-Rebuild-World-Scratch...

That risk seems like a liability to me. It's also the Achilles' heel of all the new non-fossil-fuel-based energy production methods at scale - they all end up being dependent on that same advanced industrial base and transportation network, which is complex, fragile, and incidentally still runs on fossil fuels.

> It would need to be rebooted from scratch.

With no readily-accessible source of energy, we would never be able to reboot. And everything readily-available has already been mined. We now drill deep below the ocean surface for oil.

Someone else linked a book, but this TV show, Connections by James Burke [1], follows this idea of cyclical dependencies in technology. Society starts with the plow. Definitely a great watch, even though it was made basically pre-internet.

[1] https://en.m.wikipedia.org/wiki/Connections_(TV_series)

"Dark Age America" is a fascinating book by John Michael Greer that goes into the wicked dependencies in our technology stack.

That is one of many of his books to focus on this topic.

The Ecotechnic Future, The Retro Future, The Long Decent, Not the Future we ordered - all great books on what will happen if these systems crumble from resource depletion.

As much as I want disagree with his views, they look more sound every day. He is actually doing a follow up series on these topics on his blog (Ecosophia) currently.

> I love that we have machines that make better versions of their own components, something I never even thought to think of!

This is how compilers for example are made. You initially compile the improved source code of a compiler on the previous version.

I wonder if there is an ancestry trees (or I guess chains) somewhere of the major compilers out there?

I heard an interesting anecdote about screw manufacturing. A friend of mine was in china talking to manufacturers there.

He asked about standard screw sizes and was told that there were no standard sizes or "in stock", and he should just tell them what size he wanted because their supplier just makes whatever size screws they need.

I recently read "The Perfectionists", which is a very readable book on the topic of how vital precision and machine tools was to the industrial revolution and the material world we live in. It really gets to the bottom of how the advances in our material capabilities and the invention of precision went hand in hand.

https://smile.amazon.com/dp/B072BFJB3Z/

Next: nano-tech

Femtotech! https://en.wikipedia.org/wiki/Femtotechnology

GRIIIIIIIIIFFFFITH! (Cue berserk soundtrack)

But for real, interesting read. Thanks for linking that.

Would love to see some applications of such a thing.

There was a very interesting discussion on One Hour One Life mailing list about how fast we would go from prehistoric tech to modern tech if we knew all the steps and one key point was that we would have to make several iterations of increasingly precise machinery.

I learned that in 1751, Vaucanson made this full-metal lathe: https://artsandculture.google.com/asset/metal-turning-lathe/... that was a kind of breakthrough and the ancestor of modern lathe, one of the first precision machine tool. Add that to his automata and his invention of punch cards to program looms and I am now realizing how much of a Da Vinci that guy was.

I'd rate Henry Maudslay's invention of the three plate technique for producing extremely flat surfaces (popularized by Joseph Whitworth) as the most important step toward modern mass production. All precision machining ultimately depends on accurate surface plates, so without them interchangeable parts are impossible.

See the excellent book "Foundations of Mechanical Accuracy" by Wayne Moore of the Moore Tool Company for details.

While the our industrial capabilities are indeed incredible there still is an insane complexity gap between finely threaded screws and nanometer pattering ultra-purified materials, treating them with some of the most aggressive chemicals known to mankind as a mere cleaning step. The complexity of the patterns, so intricate that people can't even design them by hand anymore. Or those barely-practical national treasures that are EUV lithography machines. And all this is mass-production.

There recently was an article on HN that fabs install microphones and other environmental sensors everywhere and use deep-learning to extract QA information from the sensors.

It's similar with nails. If you tour Jefferson's home Monticello, there was a "nailery" where the nails needed were forged by hand, one by one. Being handmade, the each nail was valuable.

People would burn scrap lumber to recover the nails.

That's wild! I thought nails were way easier because you could maybe just cast them. I guess forging would make them substantially stronger.

Casting parts like that is actually incredibly difficult. The iron will want to freeze before filling the mold cavity, since the cross section of a nail is so small. While tedious & time consuming, hand forging a nail is significantly easier than attempting to cast one. Plus, cast iron is much more difficult than forging simply due to the much higher temperatures required. A simple charcoal fire is adequate for forging, the higher temperatures required for casting necessitate specialized equipment. Hence why many village blacksmiths didn't have the capacity to produce cast iron parts.

Oh interesting! TIL, thanks!

Burn scrap nothing, In Colonial America they'd burn decrepit houses to get the nails. Enough that colonial governments passed laws against it.

Which reminds me when I was a little kid people would carefully pull the nails out of old boards and save them.

I did too. These days, I save nuts and bolts. Not because they are expensive to buy, but I don't want to spend a couple hours going to the hardware store to get one.

I thought everyone had a container where they keep a random selection of nuts and bolts that have been collected over the years...

I have two clear plastic drawers from a refrigerator I had 15 years ago almost overflowing with extra hardware. While it’s mostly just nuts, bolts and nails I also throw any extra part that comes with random widgets. I just dug through it yesterday to find two small machine screws to tie together a broken plastic connector. It’s not hoarding if you use it right?

Yeah, I have a tub with random springs, brackets, spacers, clips, and odd metal bits pulled off of devices I threw away. It's surprising how convenient it is when you need something.

Or consider how useless you would be if transported back 400 years.

You would be able to offer great advice about washing hands or the solar system generalisations. But actually create something to show people....even the simplest things would be beyond most of us.

You could draw/paint modern devices which would then eventually inspire others.

If you could figure out how to make wire, you could make a working electric motor.

You’d also need to make magnets, know their polarisation, create bearings, and have the skill to assemble all of that as well.

None of that is necessary. I built a working electric motor in cub scouts using several feet of wire, tape, some iron nails, some sheet metal cut from tin cans, a wood board to mount it on, and a battery.

Tools required: tin snips and a hammer.

A battery can be made from a jar, lemon juice, and two electrodes.

Believe me, as an 8 year old, my skills were limited to using a hammer to mash my thumb with. :-)

Electromagnets substituted for magnets, and bearings were a dent in the sheet metal that the point of a nail sat in and turned. Simple and effective.

Once you showed this to a medieval craftsman, they could reproduce it and start improving it.

I’m both impressed and have learned something new. Thank you :)

I found a variant that uses magnets:

https://www.bu.edu/gk12/tommy/Electric_motors.pdf

The version I built was far simpler, and used electromagnets instead of the magnets, but you can see the general idea.

I didn't design it, I don't know where the Den Mothers got the design from. But I found it fascinating, as you could see and feel how it worked.

I built several of them. One day, I got a little bolder and replaced the battery with an A/C cord. The motor buzzed loudly and burst into flames. I learned about alternating current that day :-)

I looked it up, people knew how to make wire then. From that you could build a working telegraph system, even a telephone system, with what was available at the time.

There is a good chance you might just end up as another Ignaz Semmelweis. The idea that washing hands would help against illnesses did not agree with the state of medical science.

(Maybe not as precise as screws, I'm not sure, but) I'm freaked out by bicycle chains, so many pieces, each manufactured to precise tolerances. $15.

Especially fun are the fact that some of them are now asymmetric with hollow pins, though that one is a bit more expensive. Incremental improvements. A bike today looks fairly similar to a bike from 50-60 years ago, but every last bit is very different.

https://bikecloset.com/product/shimano-cn-hg901-11-dura-ace-...

Near Stonehenge they found a 4000 year olds dagger. Its handle "was decorated with 140,000 tiny gold studs, each almost as fine as a human hair and set into the wood at more than 1000 to the square centimetre."

Even 21-century craftsman would find that impressive.

https://www.independent.co.uk/news/science/archaeology/stone...

Reply on news.ycombinator.com