4 ms·
More in-depth about this, with examples of the encrypted texts: https://arstechnica.com/science/2023/02/lost-and-found-code-breakers-decipher-50-letters-of-mar
by zoomablemind 4y ago
More in-depth about this, with examples of the encrypted texts:
https://arstechnica.com/science/2023/02/lost-and-found-code-breakers-decipher-50-letters-of-mary-queen-of-scots/ https://arstechnica.com/science/2023/02/lost-and-found-code-...
- mistrial9 4y agoit seems that there were plain language versions of some correspondence from Mary to her French contact using those ciphers. However the total number of glyphs is said to be nearly 150,000 ? the article says nothing about combinatorial elements of those glyphs.. it must be that they combine, to get such a big set. also interesting to see that many or most common glyphs like greek were recoded
- Aachen 4y agoFrom figure 3 in the paper linked elsewhere in the thread <https://news.ycombinator.com/item?id=34718588 https://news.ycombinator.com/item?id=34718588>, it looks like 219 distinct symbols, most with diacritics, like a plus with a dot below it, or a 4 with a dot to the right. Edit: much further down it is also mentioned "Eventually, we were able to recover the meaning of almost all the 219 distinct symbol types". Wasn't sure if I had counted correctly! It's basically a substitution cipher, with some elements mapping to the same symbol to defeat frequency analysis and a few special ones like symbols for individual place, month, or person names, or a symbol to delete or duplicate the previous character. Edit2: figure 13 gives a good overview. There's no combinatorics involved like with modern ciphers.
- jgtrosh 4y agoThere were 150,000 non-distinct symbols in the discovered letters that had to be individually mapped out.
- fho 4y ago> Mary was well-trained in the art of cipher by her mother, Marie >>>de Guise<<<, from a very young age. Really? :-)
- credit_guy 4y agoIs this a wordplay on "deguise" = disguise in French?