A little bit tangential, but if the reader is interested in steelmaking, ironmaking, and blacksmithing, then I would like to draw their attention to my business - which is hiring [1]!
There is incredible complexity involved in performing these processes on an industrial scale in the modern day. This means that many important control decisions are made using approximate heuristics. And this leads to huge inefficiencies.
For example, a typical steel plant spends many $10s million/year on energy, but the theoretical minimum needed is far less. We are working on solving this problem by achieving totally optimal control. We create data-driven predictive process models that can simulate the physics and chemistry of the very quickly. This allows us to simulate different control tactics in real time - continuously evaluating different options in response to live observations.
But the down sides should be worked out in their prototype stage, once they sell this to manufactures the process should be worked leaving only upside. If they can make even a small improvement it will be adopted just because the first that does will see the improvement in their bottom line.
What is your opinion on what Boston Metal is doing? Seem like a totally new way of refining iron and producing steel that should also be more efficient.
"Harald the Blacksmith" has a related series of articles [0] on ab-initio iron production, but with more focus on the DIY-aspects. I have used those articles a lot as reference for doing iron smelting with my scouts.
Another excellent resource on the magic of iron is the hypertext "Iron, Steel and Swords" by Helmut Föll [1] -- which is also an interesting experience just for the form of presentation.
Are we going to put every ACOUP post up on Hacker News going forward? Because I would be on board with that.
He's got a lot of good content on his site besides the iron/steel sequence, I thought his contrast between Sauron's well executed attack on Minas Tirath versus Saruman's amateurish attack on Helm's Deep was quite well done as well.
Something interesting to note is that hardening steel does not significantly change the stiffness of the material. A spring or sword that is heat treated to a soft condition (annealed) is not more floppy than the object in the normal hardened condition.
Not only does the Young's modulus hardly change between soft and hard steel, it barely changes for any steel. This provides an interesting constraint for some types of engineering. For example, if you are making a bridge you can use better steel to make it stronger, but you can't use better steel to make the bridge have less deflection for a given load. You can only change deflection by engineering the shape and design of the bridge.
For people interested in history of metallurgy from more technical side i would suggest this website: https://www.tf.uni-kiel.de/matwis/amat/iss/ - Helmut Föll: Iron, Steel and Swords - A detailed history of iron and steel from a Materials Science point of view.
Will this series go on to cover the Bessemer converter?
That changed the world. For millennia, steel was a high-cost, low-volume item, like titanium today. Then, in the 1850s, Bessemer and others finally figured out how to scale up steel making. Suddenly production went from blacksmith-scale to tens of tons per hour.
The earlier parts of the series detail the huge amount of effort and resources required to obtain workable iron; the Bessemer process alone doesn't immediately scale up your steel production, you need a vast amount of industrialisation elsewhere to support it.
"I will do my best to capture changes in metal-working techniques in the medieval period. What I am not going to cover in detail is modern steel and iron-working (that is, post-industrial-revolution), though I will occasionally note how it is different"
The old posts on the blog can be pretty interesting, though the author seems to have viewed it more as a stepping stone to becoming a political activist than as a project in its own right.
I grew up in a steel town (Port Talbot, South Wales) in the shadow of a massive steel plant. It's very interesting to be able to decompose Bret's descriptions and align each stage of the process with its modern, automated equivalent (... give or take).
The same kind of complaint was raised against the previous post and someone posted a link to the authors tweet about that. Seems like they prefer to publish the content first and then do proofreading passes.
Something I’ve learned from my own writing is that it’s possible to copy edit a piece of text into illegibility.
When I look at my stuff from the heyday of blogs, half of it is, “wow, did I really write that?” And the other half I have no idea what I was trying to say.
From the readers side I prefer that approach, too - I’d rather have the content available with some typos than stuck in the editing pipeline of the author. If it’d be a book or some professional publication, I’d probably have higher expectations, but for someone who has no dedicated person hunting copy mistakes - let’s cut them some slack.
The other thing I learned is you can get quite good at writing if you do enough of it. I've never been that much of a writer, but I ended up as the internal communications person at my company. Not exciting at all. I was actually dreading it.
After doing it for a year, I got pretty good at it (I'd like to think). I'm able to pretty quickly throw together an email or press release that's clear (avoids BS corporate speak), concise (summarize a paragraph in one sentence) and sets the right tone (why should the reader care).
Now I read others people's work (not talking about this blog post at all, I haven't read it) and subconsciously edit it my head.
My takeaway? Anyone can be a good writer with enough practice.
He's getting a lot of ripoffs copying his work verbatim, nearly all of it occurs within a few hours of being posted here.
None of it has attribution, all are selling ads.
Some are copying comments from the threads here and putting them down the bottom of their fake blogs also to give it some legitimacy.
It's a shame when people who write high quality stuff like this for free end up treated this way by digital parasites. It's not an easy problem to solve, either you lock it up and get no exposure or you get robbed.
His work will still be there years from now, the poorly proofread copies likely will be too.
I made that original comment in the last thread, it was only speculation but it's an easy, low cost strategy to deal with this.
Let it get popular on HN with no proofreading, it gets copied, then fix it in a week or two and keep the version more likely to stand the test of time.
> He's getting a lot of ripoffs copying his work verbatim, nearly all of it occurs within a few hours of being posted here.
> None of it has attribution, all are selling ads.
Unsurprising, yet sad.
Some even go to the trouble of changing a couple of words, presumably trying to cover their tracks or justify some kind of claim of "authorship":
"This week, we continue our ... stare upon pre-standard iron and steel manufacturing"
No normal English speaker would write that. Sure enough, the original had "look at", and the thief has just substituted "stare upon" without considering whether it actually reads well in the context.
If someone cared to DoS or otherwise take down that parasitical site, I'd consider it a service to humanity....
Indications of wholesale content theft or minor / poor quality repackaging of OC would be a handy thing to know. Perhaps have them removed from serps or display on feeds.
It happens enough on HN that bad re-blogs of OC get ads thrown around them. This shouldn't be.
I can recommend the book 'Men, Machines and Modern Times' by Elting E. Morison for (among other things) a fascinating look into the early development of the steel industry & manufacturing process / development of the Bessemer steel process, I never thought I'd find steel so engrossing :)
There is incredible complexity involved in performing these processes on an industrial scale in the modern day. This means that many important control decisions are made using approximate heuristics. And this leads to huge inefficiencies.
For example, a typical steel plant spends many $10s million/year on energy, but the theoretical minimum needed is far less. We are working on solving this problem by achieving totally optimal control. We create data-driven predictive process models that can simulate the physics and chemistry of the very quickly. This allows us to simulate different control tactics in real time - continuously evaluating different options in response to live observations.
[1] https://www.foresightdatamachines.com/jobs or email me: info [at] foresightdatamachines [dot] com