Path: ...!eternal-september.org!feeder3.eternal-september.org!news.eternal-september.org!eternal-september.org!.POSTED!not-for-mail From: Richard Heathfield Newsgroups: sci.crypt Subject: Re: Seriation Date: Tue, 4 Feb 2025 06:10:59 +0000 Organization: Fix this later Lines: 75 Message-ID: References: MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: 7bit Injection-Date: Tue, 04 Feb 2025 07:11:06 +0100 (CET) Injection-Info: dont-email.me; posting-host="11d224b3e045b20ace3938a9233625ef"; logging-data="1841405"; mail-complaints-to="abuse@eternal-september.org"; posting-account="U2FsdGVkX19IDHdB0gE789hOLdLott1izZ+oijAqQkjoHynJO+SH8g==" User-Agent: Mozilla Thunderbird Cancel-Lock: sha1:Vlig9+oR5bzI2V4ieSBWNXwjWyU= Content-Language: en-GB In-Reply-To: Bytes: 4678 On 04/02/2025 01:31, David Entwistle wrote: > On Mon, 3 Feb 2025 10:17:24 +0000, Richard Heathfield wrote: > >> #define UPPER "ABCDEFGHIJKLMNOPQRSTUVWXYZ" >> #define LOWER "abcdefghijklmnopqrstuvwxyz" >> #define DIGIT "0123456789" >> #define PUNCT "!\"$%^&*()_-+={}[]#~'@;:/?.>,<\\|" >> >> which isn't particularly wide, but any encryption of any of those >> characters will *always* produce a character drawn from the same set. >> When faced with something *not* from that set, SCOS leaves it unchanged. >> So I think I'm right in saying that by the above definition SCOS /is/ >> closed. > > Yes, as I recall, and as a rather sloppy adversary unaware of your > implementation, my character set ran from char(33) ! to char(126) ~. I > think it was the gap between char(95) _ and char(97) a which caused me the > most trouble. char(96) is top left on a QWERTY keyboard. I never use it, > but it gets used as an apostrophe in some text on the web. As a result > some of my checks failed to do what I expected and I didn't feel confident > posting SCOS-based ciphers. Well, blow moi down with a feather! I had no idea! The backtick does indeed fall by the wayside... and I *do* use it a lot --- as `` for "open speech marks" and \` (grave accent) in LaTeX --- so I really ought to have remembered it, but it turns out I did overlook it. And *that* means that if I use my completely insecure program to encode a LaTeX document for posting on the Net with the express intent of allowing a few people I've never met to crack the ciphertext, as well as having an easy task they'll have the additional benefit of having the backticks pre-cracked for them. In the real world it would of course be an embarrassing screw-up that hopefully would be caught by a proper test plan. In the rarified atmosphere of sci.crypt, however, I can afford to be sanguine. It now occurs to me that I could remove other characters to have them left treated untreated. For example, the following ciphertext is the (key 31 41) output from a SCOS2 variant (with just one line of code changed) that passes through everything but alphabetical characters: Byuj, ONFC hYH rroU SttW E ymXfIuw! C Qyq pf OyoZ! HKw iWNKImVS AAfi NHmcRF wggV AM wZL SlyhMwm... XZE Y *ii* dQr kk G Gys --- OV `` xvn "ZPtr lXBqdX RuAif" CEj \` (bbzJh SJypni) MG TXFFn --- xi r PrccRT ytUKL Ak SAKi kMJqnRJLnb VV, sAo SS HxjUO zui m wqa aWuwfXMx kk. gIn *sVDL* taLNH xaIQ ug y ZMn kl EFskVDHhdf EysTGNzb bSEllJK Gq VTxycS D dhPPx ssVcJqoj KIA nbUKoiQ NB wZL jpt lMMp qTF uCjaCFu ZTOomh Rx hhWOLmgO x rfm UyxnYG Z'BZ XDJhj TAE td GKiZW Uxj WRNugiZzHs, OV OlhW AH lTdFzh QS yjql Vryf dGsB'dS DlvT XAm XPEyycXLnn SKIoeWW Gm dLVxrZ bEq cQHECgPMJ vmO-BFdURAo fdV MpbY. jD ybN Prcc cJBkR LL DkfLs sY KLGsiJ vn Ya GDhVbQovkPJr sRVxE-rb Uxfn QMCgWaGvx kRMsZ ME renOEF co F JAmcGI zZcS DoSU. ey tWI KioUGyjX JRzqjVCoqS Rx zYT.CGCib, EAxUayA, G PCE gaPNFg lV xp sPRzCfZF. Yy hXU BeTaMC sc Pw AdLT X ghcIp sURIEc bVykm MGouSJPprh XH pXhF JmYV Jrhk ZMoZhHv BjeRtemMA. Rph JRjkcNv, zcO ECodVStnV GBxeQSJjrc GF vYK (Fox 31 41) cXLwqe FGsf I pOPI2 auAgNPK (CddG xxka Kye aMGm lR DEiY LFnpXKy) DgOW HhodEH xaZLGhX JPnplVyoiQ AIw SSLsaQIMqZMM smUaypvVXN: -- Richard Heathfield Email: rjh at cpax dot org dot uk "Usenet is a strange place" - dmr 29 July 1999 Sig line 4 vacant - apply within