[jbay808]

As the article explains, the most interesting thing is their seaworthy cement. Although today's Portland cement is newer, it wasn't adopted because it was superior to Roman cement, but because (until recently?) no living person knew how to make Roman cement.

Hence, we can indeed learn something by studying it.

[darkerside]

It's typically not the case that we can't make things as well as we use to. It's just that we've found ways to make them that are cheaper by an order of magnitude and last just long enough to do their job.

Roman concrete wasn't disruptive. It was disrupted.

[jbay808]

> Roman concrete wasn't disruptive. It was disrupted.

I'm amazed by this mentality. It's like the very possibly of a technology being forgotten, rather than surpassed, is axiomatically impossible.

If "everyone who knew the recipe died" counts as getting disrupted, then yes, it was disrupted. And if Roman concrete counts as more expensive because the supply is literally zero, the yes, modern concrete can be seen as cheaper.

But understand that modern Portland cement wasn't invented to improve on Roman concrete, but rather as an attempt to replicate it, because nobody until then had any guess.

[darkerside]

That's not my mentality at all. I understand technology can be lost and forgotten.

In this case, I actually think the naive mindset is the one that assumes, more durable is better. No, more durable is more durable. Is static typing "better" than dynamic typing? No, it's a different tool for a different task.

Please don't fall for clickbait. It sounds so cool to have lost a technology over time, and it certainly is possible, but I don't think that's what happened here.

Here's a quick Google source. Sorry for PDF. https://www.google.com/url?sa=t&source=web&rct=j&url=https:/...

[jbay808]

Hmm, a few points...

1) This certainly doesn't imply that Roman concrete was discarded because a more cost effective formula came along. The fact is that it was lost to civilization for over a thousand years, and then Portland cement was discovered. There was no period of time in which a contractor could choose one or the other and selected Portland cement for price or performance reasons. So beyond a doubt the Roman technology was lost.

2) That article is pretty unconvincing, being basically just a blog post by a cement guy who points out the same things that anyone else should already expect, all of which were mentioned in the comments section here. And not only because the author misspelled Raman spectroscopy.

3) Survivor bias has been addressed in several other comments. It's a valid point if someone points to a 2,000 year old bridge and says "my building needed repairs after just 10 years; clearly no Roman buildings ever had that problem". Yes, I'm sure the Romans also built buildings that fell apart within 10 years and needed repairs. Those bridges don't say anything about the quality floor. But that's really missing the point.

The quality ceiling is also not an accident which is my point about the moon landings. Most rockets launched in the 20th century were backyard fireworks, so you could point to the moon landings as "survivorship bias that doesn't reflect the crude state of 20th century rocketry in which most rockets were just backyard fireworks and few even made it to space, we just only remember the ones that did".

But that statement, while true, also makes it sound like the moon landings were a statistical fluke in which a bunch of people who had no idea what they were doing got lucky and lauched some fireworks that made it to the moon. That would just be wrong. Likewise the best Roman buildings were not built by luck, but by skilled engineers with large budgets, and they are just as impressive as they seem. It's no coincidence that the structures that survive are high budget state infrastructure projects, government and religious buildings, and estates of exquisite craftsmanship, and not random commoners' houses, even though there were probably 10,000 houses for every aqueduct.

Survivorship bias also doesn't mean there's nothing to learn from those materials. Probably not every batch of Roman concrete was perfectly formulated to perform optimally, but that's no reason not to study the surviving samples and note their impressive durability.

Randomness isn't everywhere. If you dump a thousand ball bearings in the ocean and pull them up ten years later, you expect to see a thousand rusty ball bearings. If one of them is somehow still smooth and polished, you don't conclude that erosion just acts randomly and it's survivor bias. Instead, you should look closer and see if maybe that one was actually made of 316 stainless or something.