Codebreaker Cracks 135 Year Old Message [video] | Hacker News Reader

Codebreaker Cracks 135 Year Old Message [video]

youtube.com

37 pointsmattm19 comments

Open articleView on HN

There's another article about it here:

https://www.theweathernetwork.com/en/news/science/earth-scie...

An academic paper with the solution here:

https://www.tandfonline.com/doi/full/10.1080/01611194.2023.2...

The initial blog post about the piece of paper in the dress:

https://commitmentocostumes.blogspot.com/2014/02/bennetts-br...

For me this raises as many questions as it answers in some ways. I have no idea why it would be hidden away like that.

> The pure depth of this code system is utterly fascinating. Each consonant and vowel in each word has its own meaning, so a seasoned veteran could tell the weather just by mentally breaking down each word of the telegram.

(from the first link)

This claim doesn't make sense. If the words of the telegram encoding for weather information are compositional rather than lexical, most of them shouldn't exist in English. How does it happen that all of them are extant English words? What happens if my observation of the dew point temperature needs to be encoded "luft" instead of "left"?

For pressure + temperature, the measurements were encoded in the first two consonant-vowel pairs. So, for example, in the word "diamond", the two measurements were "di" (= .26 inHg) and "mo" (= 58°). There are 50 possible pressure readings and 50 possible temperature readings, for a total of 2500 words; of these, a small fraction are encoded as "arbitrary" words when no English word could be found that matched the desired pattern. You can see a full list of all the words here: https://books.google.ca/books?id=i7YazQEACAAJ&pg=PA31#v=onep...

Keep in mind: the encoding process was meant to be more expensive to encode than to decode - each weather station only had to send out (encode) a few reports per day, but the receivers had to process (decode) hundreds.

I suspect the use of words, as opposed to just sequences of letters or numbers, was to improve robustness: it would be easier for operators to transmit recognizable words correctly (and error-correct them if need be) than if they were transmitting strings of random letters.

> I suspect the use of words, as opposed to just sequences of letters or numbers, was to improve robustness

I thought one of the articles related it to telegram pricing, which was per-word.

That's covered on page 8 of the field manual [0]. Most of the words are real, some are not, and those are called out specially. Also, there are rules for ignoring letters. Generally, the journalism glosses over much of the detail.

https://books.google.com/books?id=vICPCHgtPtkC&newbks=0&pg=P...

> Also, there are rules for ignoring letters.

This is a pretty far cry from "each consonant and vowel in each word has its own meaning". Going by the handbook, the example word of diction is 57% meaningful symbols and 43% meaningless ones. The word called for by the encoding system, diti, is in fact nonexistent, but you're free to supply whatever other letters you feel like until you've formed an English word.

The example I asked about, the dew-point indicator left, is much worse, with -- as far as I can tell -- 25% letters with independent meaning and 75% letters without a meaning of their own. The handbook notes that dew-point indicators are not part of the system, and instead a lookup table is supplied which must be memorized. The linked document does not appear to include that table, though, so it's hard to characterize the dew-point indicators.

I did appreciate the note that temperatures are to be reported modulo 100 degrees, but "the intelligence of the translators will prevent any error arising".

You're not free to supply whatever letters you want - the words are specifically prescribed. The dew point indicator words are on page 41 of the handbook linked by the parent.

I'm not sure why the dewpoint indicators had to be arbitrary, rather than using the same scheme. Maybe it was a way of providing some robustness because fields could occupy variable numbers of words, so an identifiable dew-point word would help to mark a specific field in the report.

> The dew point indicator words are on page 41 of the handbook linked by the parent.

My mistake. Pages 20-30 are all blank; I figured that was the end of the document.

dry run of some standard operating procedure for delivering messages, practice for wartime.

or she was a spy/compromised and stealing secrets from her employer that she was as clueless about as any casual observers today; she did this on the daily but she died or got arrested this particular day. maybe her husband wanted the weather for his horsetrack gambling ring or something.

for every actual important encrypted message, there are a ton of similar messages and message transports of lesser importance.

(i'm assuming it's not the case that this is the plausible deniability part of some sort of steganography scheme)

Academic paper with far more detail, including the decoding of all messages here: https://canwin-datahub.ad.umanitoba.ca/data/publication/brea...

Places where the author notes a mismatch between the “correct” word and the word on the paper, I wish they’d presented the morse code difference too. Otherwise, pretty interesting!

It's clear that the author of the pieces of paper knew several words to be incorrect: "lisstd" (should be "linnet"), "piped" (should be "biped") and "meraccous" (should be "meracious") have little x's under them (the cross under "piped" is specifically under the first "p"); while "tertal" (should be tergal), "palm" (should be "bank"), and "dsrch" (should be "duck") are all underlined.

Maybe they looked into it and found it wasn’t super relevant nor interesting, but while reading the paper I found myself wondering about the “edit distances” in those typos.

Ex: ‘e’ is dot, ‘t’ is dash; making a /t/e/ substitution seem very plausible, and my hypothesis is that including that extra layer of info would be interesting. I didn’t go through all the others, and would also be willing to believe it’s a dead end.

Related article from which the CBC YouTube video is associated:

https://news.ycombinator.com/item?id=38721406

Cool, but once again it turns out that real life is not like a Dan Brown novel. No Illuminati, no antipopes, no buried treasure, just everyday activities that apparently needed some level of protection.

Well the Illuminati wouldn't be using something you could crack, unless they wanted you to think that you had cracked something, which may or may not have been made so you think that it was or wasn't the Illuminati.

That's a great point. Or is it? Is it a joke, or are you serious? Or maybe you're hinting at something you can't say outright?

I'm going to have to use a corkboard and some red string to sort out everything that's happening here.

I guess I feel like there must be more going on though? It's not clear why a weather message would have been hidden away like that. What I read about it suggested that the weather code system was more for telegraph message "compression" reasons than secrecy, although I don't know anything about it so could be missing something.

My guess is that the message must have had some other meaning in some way, even if personal, or that there's something about putting paper in the pocket for tailoring reasons?

Maybe she just needed a place to put the message and forgot about it.

The obvious theory would point out that female clothing isn't supposed to have pockets at all. This means that anything carried in any pocket is likely to be carried in a "hidden pocket" - that's the only kind of pocket.

Reply on news.ycombinator.com