Talk:Permutation box

Cryptography: Computer science Low‑importance

	This article is within the scope of WikiProject Cryptography, a collaborative effort to improve the coverage of Cryptography on Wikipedia. If you would like to participate, please visit the project page, where you can join the discussion and see a list of open tasks.CryptographyWikipedia:WikiProject CryptographyTemplate:WikiProject CryptographyCryptography articles
Low	This article has been rated as Low-importance on the importance scale.
	This article is supported by WikiProject Computer science (assessed as Low-importance).

Confusion (pun intended)

I think this needs some documentation that the terminology of "P-box" is real and widely accepted. I'm a cryptographer, and I've never heard of this term. The entry cites reference [1], but reference [1] doesn't exist. I suspect this is a personal nomenclature that is not widely used, or something made up, or a typo for S-box. Also, the entry seems to have factual errors or confusions. It claims that it "retains diffusion" (whatever that means), but a permutation (re-ordering) of the bits cannot provide any diffusion. Overall, I think this entry should be deleted. — Preceding unsigned comment added by 128.32.153.194 (talk) 20:10, 7 January 2013 (UTC)[reply]

Bruce Schneier's Applied Cryptography uses the term in its discussion of DES and related Feistel ciphers. A little googling around convinces me that lots of others use it in this context as well. Ntsimp (talk) 23:29, 7 January 2013 (UTC)[reply]

Illustration caption: straight or expansion permutation?

The current caption of the sole illustration in the article starts with the words ”an example of a 64-bit "expansion" P-box”. Isn’t it, in fact, a textbook representation of a straight permutation? It sure looks bijective to me (and the array in the image description confirms it). — Wlgrin 00:10, 12 January 2022 (UTC)[reply]

Meaning of "permutation"

See the discussion at Talk:Substitution–permutation network#Meaning of "permutation". Dimawik (talk) 22:39, 9 October 2023 (UTC)[reply]

My correction of 'permutation box' article

Moved from User talk:Dimawik

First paragraph: The permutation does not necessarily has to be connected to the S-Boxes, but to any input-output constellation. Therefore, the entry explanation is too exclusive and therefore misleading.

Second paragraph: Although the article is about P-Boxes, it talks about S-Boxes (and P-Boxes); this is again misleading. However, this is just a secondary problem (still, it should be resolved by not talking FIRST about S-Boxes and only as a second term about P-Boxes, if the article is about P-Boxes). Then, and this is the main problem, it is written that the relation between plain-text and cipher-text shall be difficult to understand - this is correct. However, it then connects this property to the Shannon-principle of "confusion" - this is wrong, it should state "diffusion" (this is the second Shannon-principle). ('Confusion' - the first Shannon-principle - is provided by S-Boxes and it breaks the connection between the encryption/description-key and the cipher text.)

Thus, there are major mistakes in this article and I show the way how to correct this. The source for my statements can be seen here: https://en.wikipedia.org/wiki/Confusion_and_diffusion

--> But attention: This article is correct in the sections 'Definition' and 'Theory'. The second initial paragraph has some mistakes too - unfortunately. However, I will correct this later. Let's agree on the initial problem first. Unicornbrain (talk) 04:37, 19 November 2024 (UTC)[reply]

(1) Many cipher designers avoid both "P-Box" and "permutation box" terms, using instead the "permutation layer". We can handle it in few ways: either create a redirect Permutation layer to point here, or change the title of this article. Yet another option is to WP:merge this WP:stub into Confusion and diffusion#Diffusion. What do you think? Dimawik (talk) 20:54, 19 November 2024 (UTC)[reply]

(2) I do not understand your "any input-output constellation" terminology. In my mind, the P-boxes (for the avoidance of doubt, I dislike this abbreviation, but it saves time typing here on the talk page) are only well-defined for the substitution-permutation network where they are neatly slotted between the substitution layers, so our current text is reasonably correct. Could you please provide an example of the cipher with P-boxes connected to the constellations, so I can understand the issue? Dimawik (talk) 21:06, 19 November 2024 (UTC)[reply]

(3) In my mind, there is no easy way to discuss permutation without substitution (and vice versa), so a mention of substitution is inevitable. I agree with you that the use of S-box here without any introduction is very confusing. I took a liberty to change the offending sentence so that proper context is just one click away. Feel free to fix it to your liking or undo my changes and start afresh. Dimawik (talk) 21:20, 19 November 2024 (UTC)[reply]

(4) I totally agree with your reasoning about the confusion and diffusion and hope that my change described at #3 resolves your concern. If not, feel free to change the text. It seems that we can easily reach consensus if the discussion is split into small chunks (thus my multiple individual replies). Dimawik (talk) 21:27, 19 November 2024 (UTC)[reply]

I think your changes are ok - especially the mentioning of diffusion right at the beginning; at the same time, they are a good example of the subjectivity of "what is difficult to understand". I agree that a substitution-permutation network needs both explanations (S... and P...). However, my initial intention was to correct an obvious mistake in the text. And this mistake is corrected with the mentioning of "diffusion". Unicornbrain (talk) 06:47, 21 November 2024 (UTC)[reply]