News for hackers. Stuff that matters.
Re:The Zodiac Killer's most uncrackable cipher has, at last,... by meowface on Friday December 11, 2020 @08:07PM (#25394091)
He does consistently misspell it as "paradice" across multiple letters, but why is it necessarily intentional? Two potential explanations for the consistency are that he intended to call attention to it, or that he simply thought or assumed the word was spelled that way. It's a reasonable phonetic spelling, using an existing English word ("dice").

He has a lot of other phonetic misspellings like "victoms", "darck", "epasode", "pencel", "paterned", "cruse", "coupples", "untill", "allready", "dissapeared", "rubed", "positivily", "ventalate", "silowets", "choaked", "extreamly", "posibly", "twich", "comitt", "descise", "saterical", "comidy", "butons", "considerbly", "promiced", "shure", "inthusiastic", "phraises", "nucenses", "truley", and mistakes like "to many" and "loose" for "lose".

It's possible there's a meaning to it, but I suspect the most boring explanation is the most likely one, there. He may have spent a lot more time speaking than reading/writing and perhaps might not have finished high school, or maybe just happens to not have been very good at memorizing spellings.

His misspellings sound a lot like phonetic misspellings I still regularly see online today, and like how people tended to write English more phonetically and less consistently a few centuries ago. We take it for granted, but English really is not at all an intuitive language when determining spelling from pronunciation or vice versa, compared to more consistent languages like German.

He does seem to have an interesting affinity for using the letter "c" almost any time a soft "c" or hard "c" could be called for, and sometimes eliding other letters. Like "twich", "darck", "promiced", and "paradice". I wonder if an automated analysis of basically every piece of writing you can find from around that time and area, taking some of those patterns into account, could uncover something. Like every single physical letter written to every newspaper about anything. It'd probably require way too much manual digitization labor to be feasible, though.

Re:The Zodiac Killer's most uncrackable cipher has, at last,... by asveikau on Friday December 11, 2020 @09:58PM (#25394970)
Whenever I've read material from the '60s or '70s, I've often noticed that phonetic misspellings were much more common in print in that time period than they are today. I always thought this was because it was much more tedious to check spelling before word processing software became common.

Typos on internet forums notwithstanding, I might even guess the age of auto-correct, spell check, and, even things like how some OSs put squiggles over misspelled words in every text field (I think Mac was the first place I noticed this -- though I'm running firefox on FreeBSD and it's doing that in this edit field right now ...), has made the average literate person a better speller than that era. We get so much realtime feedback on spelling when writing on our devices, these opportunities to learn to spell, notice and correct our mistakes would not have at all been so accessible in 1969.

Re:The Zodiac Killer's most uncrackable cipher has, at last,... by meowface on Friday December 11, 2020 @10:35PM (#25395239)
Definitely. There's also probably a bit of selection bias when seeing what HN users think of it; a high percentage of users likely have (at least) decent writing skills and probably read things fairly often (even if it's just blog posts/emails/chats).

If you read/write very rarely and speak very often, these kinds of misspellings may seem pretty natural. But if you write a lot - perhaps even a lot more than you speak - they may stick out like a sore thumb and your brain might flag them as highly unnatural and a potential sign of hidden meaning. (I'm probably like 99.5% writing 0.5% speaking, myself, and I bet many people reading this are, too.)

Then combine it with one of the most mysterious and analyzed crime stories in US history, with a cryptic, megalomaniacal serial killer who loved to use ciphers and symbology to toy with the press, public, and police. I can definitely see why people might suspect something deeper to the spelling, but the null hypothesis always needs to be carefully considered, first.

And a side note: does anyone know what the trends over the decades are like for less "hodgepodge-y" languages, like Spanish and German? They seem to have rules for pronouncing based on spelling and spelling based on pronunciation that are a lot more predictable and consistent than English's. I would guess that the popularization of spellcheck/autocorrect might be less significant for such languages, though I'd also love to be proven wrong.

Re:The Zodiac Killer's most uncrackable cipher has, at last,... by asveikau on Friday December 11, 2020 @11:08PM (#25395430)
> And a side note: does anyone know what the trends over the decades are like for less "hodgepodge-y" languages, like Spanish and German?

I'm an amateur observer of this, but speaking in centuries rather than decades I can tell you that Old Spanish up until a little bit after the time of Columbus had different phonetics for several consonants, and thus different spelling. eg. There was a /ʃ/ sound spelled with X, as well as a /ʒ/ which was spelled with J or GI, GE -- this all merged and is today spelt with J or GI, GE, pronounced [x]. Also what is today /θ/ in Spain was once distinct /ts/ and /dz/, written as ç and z respectively (today c and z). /b/ and /v/ had not yet merged and their spellings have since been sort of "re-based" to look like their Latin roots. <h> for a time was not silent, and many words that start with <h> used to have an /f/.

So, the interesting thing about this relative to English: when pronunciation changed around 1500, they also updated the spelling.

Re:The Zodiac Killer's most uncrackable cipher has, at last,... by bernawil on Saturday December 12, 2020 @04:50AM (#25397108)
> <h> for a time was not silent, and many words that start with <h> used to have an /f/.

Probably related, most words with an H in Spanish use an F in Portuguese. Like 'HONGO' - > 'FUNGO'.

In the 17th century there was a concerned effort to make the written language consistent in writing.

Re:The Zodiac Killer's most uncrackable cipher has, at last,... by asveikau on Saturday December 12, 2020 @09:35AM (#25398441)
Yes, and go further back and those words in Latin all have f.

Eg. hongo from fungus. hacer from facere. hijo, hoja from filius, folia (/lj/ to /x/ being another change relative to Latin). Too many other examples to list.

Re:The Zodiac Killer's most uncrackable cipher has, at last,... by xenadu02 on Saturday December 12, 2020 @02:54AM (#25396565)
How late was it that there was still wide disagreement or variance in spelling?

Even when the printing press arrived and dictionaries were published most people who knew how to read/write outside high society, monks, scholars, etc would have used phonetic spellings or whatever spelling was used in their area for a very long time.

Spellcheck has done more to freeze spelling than anything since the Dictionary. I don't know if that matters but I do find it interesting. In the past if a prominent newspaper or popular writer decided to spell a word differently it could catch on. Now the spell checker pushes back against such changes.

Re:The Zodiac Killer's most uncrackable cipher has, at last,... by asveikau on Saturday December 12, 2020 @01:27PM (#25400309)
There are still a few where variations are excepted. I always think of "cancelled" or "labelled" which some people spell with one L near the end.

Sometimes that's a question of which country, US usage, UK usage, whatever, but sometimes multiple spellings are found in either place.

Re:The Zodiac Killer's most uncrackable cipher has, at last,... by weare138 on Saturday December 12, 2020 @03:47AM (#25396825)
Another explanation that makes more sense to me is it was done intentionally to make it harder to decipher. When I was young our newspaper used to have cryptogram puzzles in them based on a simple substitute cipher. The first thing you look for is repeating letter patterns like double letters and common word patterns like words ending in 'ing' or 'tion' for example. In his cipher he tends to drop a letter in words like 'button' and adds double letters to words that don't have them. Plus in his plain text letters to the press and police he tends to spell everything correctly.
Re:The Zodiac Killer's most uncrackable cipher has, at last,... by jrumbut on Friday December 11, 2020 @11:28PM (#25395550)
I always assumed it was to make deciphering just a little more difficult by having correct solutions include nonsense words and mess with the distribution of different letters.
Re:The Zodiac Killer's most uncrackable cipher has, at last,... by meowface on Saturday December 12, 2020 @04:41AM (#25397070)
Most of the time, he appears to have also misspelled various things in his many plaintext letters, and generally the same things consistently, shared between encrypted and plaintext writing. They had the opposite effect of hardening the cipher and actually served as perfect "cribs" to aid decryption [1].

There are some occasions where a word is spelled correctly in one place and incorrectly in another, but this isn't necessarily anomalous for people who don't write or read very often or don't have a lot of spelling memorized. In some cases they might even be using some words they've only heard and spoken and never seen before (or only decades prior), and then might see the correct spelling later, or look it up.

There's also the possibility he did it consistently because he wanted to mislead police about his knowledge/education/region. Haven't dug into it or anything, but I could definitely see either one being true (genuine mistake vs. subterfuge).

[1] https://en.wikipedia.org/wiki/Crib_(cryptanalysis)

Re:The Zodiac Killer's most uncrackable cipher has, at last,... by blindm on Saturday December 12, 2020 @01:54PM (#25400506)
> He does consistently misspell it as "paradice" across multiple letters

Yes but I wouldn't read into it that much. Remember, reading the deciphered ramblings of a deranged psychopath is obviously going to have misspellings and weird idiosyncratic oddities in them.

My take is (if the killer was in any way rational) is that he knew the plaintext would be unraveled at some stage. If not by quantum computers, then a team of dedicated sleuths who know how to program. So yes: the misspelling could be deliberate.