Talk:Substitution cipher

Block cipher substitution

From article: Modern block ciphers such as DES and Rijndael can be viewed as substitution ciphers on a large alphabet. They treat each 64-bit or 128-bit block of the plaintext as a symbol and perform several rounds of substitutions and transpositions on the bits in the block to effect a general block-to-block substitution. The various block cipher modes of operation are analogous to the various polyalphabetics, while "randomized encryption" is similar to a homophonic substitution.

Matt, I would argue against your statement here that ECB block cyphers are substitution cyphers. While true from a sufficiently advanced level of abstraction, this is too advanced a level of abstraction for the reader we are justified in expecting for this article. Perhaps as an 'advanced perspecitive' note? ww 18:43, 21 Apr 2004 (UTC)

Well, while its wording is not particularly helpful, I don't think it's too advanced or abstract a concept, especially placed right after polygraphic substitution; all that's changed is the alphabet. I've seen cryptography tutorials introduce first classical ciphers, then introduce block ciphers using exactly sort of exposition. (It also wasn't originally my statement, I just sharpened it a little). — Matt 19:31, 21 Apr 2004 (UTC)

Matt, Can't agree. Should I just edit to try to accomodate and see if you can live with it? I agree that the wording seems reasonable, I just think it's confusing due to LoA issues. Thoughts? ww 20:28, 21 Apr 2004 (UTC)

Well, maybe we could move it to block cipher and just include one sentence in this article like "Some more modern ciphers can be seen as polygraphic substitution ciphers on a binary alphabet; see Block cipher and Substitution box for more detail." ? — Matt 20:43, 21 Apr 2004 (UTC)

Matt, Seems reasonable. But the issue is important enough (evidence the original edit and your endoresment) that perhaps more would be well. A comment along the lines of '...depending on the level of abstraction, it is also possible to see modern (ie, bit oriented) cyphers as ..."? This has the virtue of saying _how_ it is that they can be seen that way, without adding likely confusing discussion here. Not just a bald statement that it's possible. In block cypher (or perhaps better in substitution box) we would add a para noting the more general perspective viewpoint. This puts the 'hard stuff' in a technical article which is a virtue, I suspect. A good thing. Shall we? Shall I, will you? ww 16:46, 22 Apr 2004 (UTC)

Go for it; I can fiddle with what you end up with. (BTW, the stuff about homophonic and randomised encryption in that para looks less useful to me). — Matt 17:25, 22 Apr 2004 (UTC)

NB -- right shift! all the way!

Matt, Went for it. Added a couple of paras on mechanical substitution, and a couple on the one time pad (which might not strictly need to be here, but... Comments?), and cleaned up the modern section and an analogous couple of places elsewhere. Plus multiple copyedits. It's tighter now.

Save some over mathematization (in discussing the Hill cypher) I think this is pretty good. It's complete enough, and has enough links to be helpful. After you tweak (I'm sure I count on you for that!), do you agree as to quality? Perhaps this could be a feature cand? Thoughts? ww 19:35, 22 Apr 2004 (UTC)

Looks like some good work; I'll add a few comments tomorrow, but I'm a bit tired tonight to go through it carefully enough! — Matt 20:34, 22 Apr 2004 (UTC)

cypher changes

Matt, Half the article was cipher, half cypher. I had to change one, so... Probably didn't get them all, though. I wish there could be a settlement of this. Maybe something along the lines of count the instances and that which has more spellings in an article sets the standard for that article's life. Couldn't work though.... ww 19:37, 13 May 2004 (UTC)

I think it seemed particularly odd in this case because the title of the article is spelled "cipher", and so to spell it differently throughout the article would be strange. I agree, though, that articles should standardise on a single spelling, and not mix. Note, however, that this article didn't contain a single "cypher" spelling until your series of edits at the end of last month, see: http://en.wikipedia.org/w/wiki.phtml?title=Substitution_cipher&diff=0&oldid=3296465

— Matt 08:39, 14 May 2004 (UTC)

Matt, I haven't figured out what my position is on article titles having cipher and content containing cypher or vice versa. I notice it sometimes, and not others and am annoyed when I notice discrepancies, as I'm with jon on preference in this question of spelling, 'y' v 'i'. I'm certainly opposed to the Miss Fidditchs who enforce (their idea of) 'correct spelling' in such cases of ambiguity. This language has, for reasons opaque to me save the historical origin of the 'problem', taken the position that a foolish consistency (ie, in re spelling) is the hobgoblin of small minds. But this reifies language and is impermissible. I don't understand what's going on, save as I noted the historical development.

(ww, as an aside: I don't suppose I could ask you to simplify some of your comment wording sometimes, for the sake of small minds such as myself who struggle to understand various phrasings (e.g. "Miss Fidditch", hobgoblins, reifies...))

Sorry. No complication intended. The hobgoblin business was a joke; it's a famous (on this side of the pond, anyway) phrase from Emerson used almost exclusively to mock pedantry and pedants. Miss Fidditch is less of a joke; she is the metasyntatic variable placeholder for the picky prune English teacher everyone had in 6th grade who took prescriptivism waaay too seriously. Her mind is one of those beset by hobgoblins.

The reifies comment was not a joke, though perhaps less momentous in re English (spelling) than in other contexts. Reification is the process of treating as an actual thing (eg, with properties and perhaps even with motives) an abstract concept. It is, in my view, the single most pernicious reasoning fallacy humans commit, and perhaps that which most reduces to nullity our (limited in any case) claim to sapiens. Thus, there is a concept -- the country 'Fredonia'. I reify if I ascribe to 'Freedonians' properties such as lazy, intelligent, dangerous, beneficient, ... Indeed, I reify merely by using such a phrase as Fredonians. The Nazis and others did this to extremes, racism generally is a reification fallacy, most political controversy is reification somewhere underneath, and so on and dismally on.

It was the fartherest thing from my intent to include you in the 'small minds' category. Neither figuratively, nor humourously. After all, you understand block cypher innards! ww 16:48, 14 May 2004 (UTC)

With respect to cypher vs cipher, I think that we should consistently stick to one spelling within an article; this is analogous with the Wikipedia's style policy on American vs British English. (Note that I do not consider cipher vs cypher to be American vs British English). I also think that the spelling should match the title spelling. This could be achieved by either A) moving an article to rename it; or B) Changing the spelling within an article. — Matt 15:11, 14 May 2004 (UTC)

Since title space is a special name space, how to manage has left me in au uncollapsed superimposed state.

As for my edit changes adding instances of cypher, I wasn't noticing that issue while I was attmepting to tighten it up and resolve the "bit twiddling = substitution cypher" conceptual quicksand. I was in respect to spelling, on auto pilot. And didn't realize I'd been the source of some cy in an article sea of ci when I returned to it with an eye to proposing it for the possible FA list. ww 13:58, 14 May 2004 (UTC)

To do

• Merge most of the polyalphabetic section into Polyalphabetic cipher; leave one or two summary paragraphs here.

— Matt 03:40, 23 Jul 2004 (UTC)

• • Hmm. Wikipedia is a wonderful thing, but IMHO it has quite a few defects. One of them at present is the tendency to wikify off every little aspect of a subject so that instead of structured articles we end up with scores of snippets. (I am probably as bad as anyone with this, but I try to avoid it.) This is highly amenable to spending vast amounts of time surfing from link to link (and probably very costly to bots, for that matter), but fairly poor for didactic purposes. One ends up with a tree structure of links open in ones tabbed browser, and not finishing the exploration of even a single concrete concept before smoko is over. I confess I feel a certain parental attachment to the current structure of the article because I created most of the layout [1], with the specific intention that it be didactic. At the present time, the article contains 17 kB (or ~2400 words) which I think is a good size - it can be easily read at a sitting, but contains a nice overview of the subject. Rather than merging the 600-odd words of the polyalphabetic section into the polyalphabetic page, I think it would be better to keep that here as a reasonably accessible overview, yet expand (no doubt gradually) the polyalphabetic article into something more detailed and sophisticated: pushing it up to ~20 kB, with more history, more on general attacks, and more examples. Then the articles on specific examples can be viewed as the third (and presumably final) layer of a structured syllabus. That was what I had in mind, although at my present rate it will take a very long time indeed... --Securiger 09:40, 23 Jul 2004 (UTC)
    • Sure, especially if the polyalphabetic cipher article has room for expansion. Usually, if there's two treatments of a topic "Foo,", one in [[Foo]], and one as a subsection of [[Bar#Foo]], and both roughly the same length, then it's a sign that refactoring might be appropriate...I guess writing more is also a solution! If we could get the "polyalphabetics" article up to 20K it would be awesome — both this article and polyalphabetic cipher are in the provisional WikiReader Cryptography table of contents. — Matt 17:10, 23 Jul 2004 (UTC)
  • Incidentally, I've also been thinking about an illustration. Being fairly abstract, it isn't easy to think of an instructive illustration, but perhaps a decorative, eye-catching one like this crown copyright image of Mary's nomenclator? -- Securiger 09:50, 23 Jul 2004 (UTC)
    • I agree; there's bound to be quite a few suitable historical illustrations. Wouldn't the Mary nomenclator illustration be in the public domain, given that it was created hundreds of years ago? — Matt 17:10, 23 Jul 2004 (UTC)

Voynich manuscript?

"More artistically, though not necessarily more securely, some homophonic ciphers employed wholly invented alphabets of fanciful symbols. (See Poe's The Gold Bug for a literary example; cf. the Voynich manuscript.)" Is there any good evidence that the Voynich manuscript uses a homophonic cipher? If not, it should be removed as an example. (I'm aware that it's been postulated that the Voynich MS is a homophonic cipher, but just about every type of cryptography has been postulated with respect to the Voynich MS.) Chuck 22:52, 12 January 2006 (UTC)

And the Gold Bug isn't an example of a homophonic cipher either; I've removed the entire parenthesis. Thanks for spotting that; I remember a decent PD image in David Kahn's The Codebreakers that would illustrate this well. I'll try and get round to scanning it in. Matt Crypto 09:30, 13 January 2006 (UTC)

Porta's Diagrammatic Digraph

"In the same De Furtivis Literarum Notis mentioned above, della Porta actually proposed such a system, with a 26 x 26 tableau filled with 676 unique glyphs."

A small point, but della Porta used only 20 letters in Latin and Italian. --Steve 04:29, 4 June 2006 (UTC)

You're right, well spotted. I corrected it, thank you. (Note that you can edit articles!) -- Securiger 09:23, 7 June 2006 (UTC)

We need a list

We definitly need to make a list of ciphers.

Homophonic substitution is monoalphabetic

According to The Code Book, homophonic substitution is monoalphabetic because one letter in the cipher can only match to a single letter in the plaintext, even though a single letter in the plaintext can be matched to multiple letters in the ciphertext. (oh, and I accidentally hit enter too soon when typing the description of my change) --Sydius (talk) 00:14, 19 August 2008 (UTC)

Nomenclatur homophonic?

I think nomenclator isn't really homophonic, so I split the section and gave nomenclator its own section. --Sydius (talk) 00:23, 19 August 2008 (UTC)

I put it back to the way it is, since I'm not 100% sure. --Sydius (talk) 00:40, 19 August 2008 (UTC)

Sales catalogs?

This needs more context explaining what is encrypted or who does/n't get the keys or why. Without that it's an odd detail that doesn't belong in this article. —Preceding unsigned comment added by 173.88.143.41 (talk) 14:49, 18 April 2009 (UTC)

Nomenclator

I think the nomenclator in ancient Rome was the slave which walked beside the master in the street, remembering him in a low voice the name of the people they casually met. So, the master could greet other people by name, as if he spontaneously remebered their names. It was particularly useful for political candidates. Lele giannoni (talk) 13:52, 10 August 2011 (UTC)

<span> revert war

Hi. Just stumbled across this page, in its current state. I'd be curious to know why this anchor is wanted here.

I'm not an expert in WP markup, but I've found a fix that seems to allow the anchor without breaking the page.

As was, the changes were breaking the page, at least for me: I was getting a dashed box around the text "== Simple substitution==" (just like the boxes around the messages further down). I can provide a screenshot if desired. But it looked roughly like this:

== Simple substitution==

The page linked to (Wikipedia:ANCHOR#Section_linking_.28anchors.29) mentions that a terminating </span> is needed somewhere.

I've changed to the {{Anchor}} syntax anyway for neatness, but even so it appears headings only like to go on their own line. So I put a linebreak after the anchor, and forced the TOC to appear before it (otherwise it would go directly before the heading, and after the anchor.)

CountingPine (talk) 22:40, 14 August 2012 (UTC)

>> Thank you. Your edit works for me.

jepler (talk) 00:42, 15 August 2012 (UTC)

Potential Corrections

In the Simple Substitution section, shouldn't the Ciphertext alphabet be ZEBRASTUVWXYCDFGHIJKLMNOPQ instead of the given ZEBRASCDFGHIJKLMNOPQTUVWXY? I could be mistaken, but I thought the ciphertext alphabet followed the remainder of the standard alphabet from the end of the key and then starts over at A (or the first unused letter). Orca239 (talk) 01:53, 10 November 2012 (UTC)

Futurama "Alien Langugage 2"

The description given in the text wasn't quite accurate. It said:

Later, the producers created a second alien language that used a combination of replacement and mathematical Ciphers. Once the English letter of the alien language is deciphered, then the numerical value of that letter (1 through 26 respectively) is then added to the value of the previous letter showing the actual intended letter.

Adding 1 through 26 would mean that two repeated ciphertext symbols meant that the second represents a "Z", but in fact it means the second represents an "A". You need to map the letters to 0 through 25, not 1 through 26.

I fixed it, but it may not be clear as written to someone who doesn't think like a programmer, so someone may want to go over it. --50.0.128.185 (talk) 03:08, 19 March 2016 (UTC)