Talk: Encryption
From open-encyclopedia.com - the free encyclopedia.
| Contents |
articles should be merged, too similar
Encryption and Cipher are currently too similar. They should be merged, or Cipher should be specialized to the customary (though vague) subset of private-key encryption.
disagree, and problems w/ diagram
Tromer, Your observation is a common misunderstanding. Encryption can be done in several ways, only one of which uses a cypher. As for 'cipher' being customarily a subset of private-key encryption, that is true only if by private key one means symmetric key cypher. At least that's the sense I take from your comment. The spelling is (cy v ci) is irrelevant as to meaning, but excites comment (see Talk:Cryptography for some history on WP). As the diagram notes, cyphers come in symmetric and asymmetric flavors, and symmetric cyphers are sometimes (confusingly) called private key cyphers. Asymmetric key cyphers are sometimes (wrongly) taken to be all public key/private key cyphers. Not all are, as there exist some in which there are no public keys.
As for the diagram, I will note that the placement of rotor machine is probably incorrect. Rotor machines (as the Hebern machine, Enigma machine, SIGABA, Typex, and even the non-rotor Japanese stepping switch machines) are fundamentally substitution cyphers, albeit polyalphabetic ones. The diagram should reflect this. ww 15:12, 19 Jul 2004 (UTC)
I'm going to make a text version of the diagram here, so it's easier for me (and presumably others) to edit:
- Ciphers
- Classical
- Substitution
- Rotor machines
- Transposition
- Substitution
- Modern
- Public Key (shouldn't this be "assymetric key" ?)
- Private Key (shouldn't this be "symmetric key" ?)
- Stream
- Block
Do "Stream" and "Block" really only apply to "Private Key" ciphers ?
- Steam and block: in usage, yes. "Public key" and "private key" are, essentially, synonymous with "assymetric" and "symmetric key".— Matt 09:18, 22 Aug 2004 (UTC)
Taxonomy approach
See also Image talk:Cipher-taxonomy.png. There is a problem with making a taxonomy of ciphers. There are (at least) two ways of approaching it, and each is somewhat unsatisfactory. The first is a more "abstract" classification, dividing the ciphers strictly according to how they function. The problem with this approach is that you then make distinctions that are never made in practice: the distinction between substitution and transposition is only really used in the context of classical ciphers; the distinction between symmetric and asymmetric ciphers is only really used in the context of modern cryptography, and so on.
The alternative approach, employed here, is to divide ciphers into sections according to how they are divided in practice. The problem with the "usage-reflecting" style is that, e.g., a classical substitution cipher isn't labelled as a symmetric key cipher, which might be desirable, but I think it's more important to reflect usage. — Matt 09:18, 22 Aug 2004 (UTC)
Encryption: hash functions?
I moved this recent addition here temporarily:
Cryptographic hashes, also known as one-way hashes or message digests, are used to encrypt data so that it cannot ever be decrypted, but it can be recognized because the same data always produces the same output. Other unique features of this form of encryption include that no matter the size of the input, the size of the output is always the same (the size of the output varies from algorithm to algorithm), and that no keys are used. Popular algorithms include MD5 and SHA. See the Cryptographic hash function entry for more information.
I'm not sure this is the right place to diversify into a discussion of crypto hash functions -- secure hashing isn't usually described as "encryption", although it's certainly part of cryptography and symmetric key cryptography. — Matt 13:14, 2 Oct 2004 (UTC)
Cryptographic hash functions
While cryptographic hash functions are certainly not a cipher, they are a vital part of encryption technology. Someone unfamiliar with the plumbing of encryption will just type "encryption" into Wikipedia and hope to get an overview of the subject with links to the details. This seems like a reasonable approach to making the Encryption and Cipher entries different so they work best for users. Encryption is a more general term.
It is logical that the Encryption entry should have brief descriptions of symmetric and asymmetric ciphers, hash functions and encryption-strength pRNGs in order to best serve users who may be unfamiliar with the categories or taxonomy related to the subject.
Vancegloster 21:27, 3 Oct 2004 (UTC)
- OK, I can see what you're getting at here; you're saying that "encryption" is sometimes used to mean "encryption technology", which is essentially the field of "cryptography". Currently, we have an overview of the field in Cryptography, which deals (hopefully) with everything, including the entire gamut of PRNGs, digital signatures, hash functions, secret sharing schemes, authentication protocols, MACs and so on. This article, which deals with ciphers, is termed Encryption — in my experience, most cryptographers use "encryption" to refer to the action of ciphers, so it's a good title. You've pointed out, though, that sometimes people are looking for an overview of the entire field when they search for "encryption". Your suggestion is to move the cipher stuff into an article called Cipher, and have Encryption survey all the cryptography mechanisms. The problem with this is that most of Cryptography and Encryption would then be duplicating each other to a large extent. I'd propose an alternative: to this page, add a "disambiguation" header saying something like:
- This article is about algorithms for encryption and decryption. For an overview of cryptographic technology related to encryption, see cryptography.
- Do you think that would suffice? — Matt 22:14, 3 Oct 2004 (UTC)
This does seem reasonable. Vancegloster 23:54, 7 Oct 2004 (UTC)
