blackstache

Sorry to hear this happened as this shouldn't have been lost. I'll make sure we get in touch with you.

Hey, that was me! What a pleasant surprise, you made my day. Have a good one yourself!

Yes, definitely. Our research group and some others have been actively working with and scanning the materials from this collection to advance our understanding and methods. This is focusing on both open fragments and intact, sealed scrolls. Both are helping us work towards ultimately being able to read the sealed scrolls through CT.

Ultimately, the sky is the limit with the Herculaneum collection. It will probably be a slow and steady build though. We are working to prove the methods on the real Herculaneum material, at which point we can start extending the method to more of the intact scrolls. The segmentation and some other steps of the process still involve some manual work, so even once the concept is proven we will want to further automate those steps. If this all goes well, it would be an incentive for those in charge of the archaeological site to further explore parts of the still-buried Villa and look for more scrolls from the library. But even of those already excavated, we are looking at on the order of hundreds of intact scrolls that contain many columns of text each and are currently entirely unseen.

The technique definitely applies to other artifacts! The core pipeline we call Virtual Unwrapping and has been used to successfully reveal some ancient texts[1]. The primary challenge in this post is addressing the "carbon ink problem" specifically, where the ink looks identical in density to the uninked papyrus. But some artifacts make it easier, for example they are written with iron gall ink which shows up quite clearly in X-ray CT.

[1] https://advances.sciencemag.org/content/2/9/e1601247.full

Oh, cool to see this here! I am one of the authors of the paper, I can answer questions later today.

If the morphology hypothesis (of the above paper) is more or less true, then it would indicate it is possible to detect the ink signal if the resolution of the CT scan is roughly at least as small as the thickness of the ink layer on the papyrus surface. We have measured this on some proxy manuscripts made in the lab to be on the order of 5 microns. So if you can scan an entire scroll at that resolution, in theory the data is all there.

The micro-CT machines currently available are either optimized for smaller object sizes or for lower resolutions. The challenge is to build a machine that can scan a "large" object such as an entire scroll at a high resolution. The above comment is saying that you could design and build such a machine using components currently available (expensive, but possible).

That's not the sole obstacle though, next you have to extract the information you want (visible text) from the resulting data. There are some more challenges here yet to be solved, notably segmenting out the individual layers from each other. This is particularly hard with the Herculaneum scrolls, as there are often hundreds of layers all smashed together, and they are wrinkled and warped. But we're working on this too!

Edit: as for the contents, it could be just about anything from the time period. Many of these scrolls have been opened manually over the last two hundred years since their discovery. This has largely destroyed them into thousands of fragments, but in many cases has revealed some text, so we know some of the contents of the library. Much of it is by Philodemus, who was an Epicurean philosopher but not particularly notable. It is believed that there could be more scrolls from other sections of the library still buried, as much of the town is not excavated. If it is shown that we can indeed read the contents noninvasively, it might be an incentive to dig up more of them and see what is there.

I can second the recommendation of Herculaneum, it is an incredible site and I really enjoyed my visit.

There was a collection of papyrus scrolls in the Villa dei Papiri there, which were carbonized by the ash. They were damaged and preserved at the same time by the eruption, so we still have many of them today but the intact scrolls fall apart if you try to open them by hand (similar to trying to unroll a piece of charcoal).

I have been working for a while on a research team that is attempting to read these scrolls noninvasively using micro-CT. The materials of this particular collection are the perfect storm of challenges for this approach, but we are chipping away at the problem. It has been quite a lot of fun to work on. Should anyone find it interesting, here[1] is some of our work from earlier this year, showing a proof of concept on how we can distinguish carbon ink from carbonized papyrus in X-ray even though they appear identical at first.

[1] https://journals.plos.org/plosone/article?id=10.1371/journal...

Bret Victor has an exploration of this question[0] (2015) that I have found helpful.

[0]: http://worrydream.com/ClimateChange/

We likely agree on the larger sentiment here, but in fairness to the author I think the context of the post makes this more acceptable. The textbook recommendations are all submitted in the comments below the compiled list, and the author's own submission[0] includes at least a disclaimer and a justification for why the author felt the need to write the book.

0: https://www.lesswrong.com/posts/xg3hXCYQPJkwHyik2/the-best-t...

Edit: I have not read this book or any others in the domain, just wanted to add some context I did not already see in this thread.

This might give the beginnings of an idea of what would happen:

https://what-if.xkcd.com/1/

This is just like http://windirstat.info/ which can be run on the desktop. Very handy tool...