From: Ken M. <ken...@gm...> - 2005-07-05 18:12:20
|
Hi, all. Looks like activity in this project is pretty light, but i'm hoping some of the folks knowledgable about mailcrypt's design are reachable, and can give me some guidance. I'm incorporating topic encryption into my emacs outliner, 'allout', and would like to include both symmetric (which maybe mailcrypt refers to as "conventional") and key-pair pgp/gpg modes. I'm a novice when it comes to public key encryption, and may be making things harder on myself than they need to be. In fact, for personal, day-to-day uses like journal entries and such, i see symmetric encryption (enhanced with key verification and hinting) as being the default mode - key-pair being somewhat heavyweight, ideally having a difficult passphrase and choice restricted to only established keys. It's quite possible that thinking is misguided, something suggested by the lack of provision, and even apparent obstacles, that mailcrypt facilities pose to doing symmetric encryption. Mailcrypt does provide nicely for symmetric *decryption*, which further leads me to suspect that the impedence against symmetric key encryption is a deliberate design choice. I could do symmetric encryption with crypt++, using mailcrypt for key-pair encryption and all decryption, but would love to not depend on both packages. First, though, i wanted to understand the situation better. So i'm asking: - whether i've missed something, and i can do symmetric-key pgp/gpg encryption with mailcrypt without major contortions - if not, whether it's so hard on purpose, and if so, why. - If it is hard, but not on purpose, would any of you be willing to work with me to make it easier to do symmetric-key encryption? (the machinery would be pretty trivial, but keeping consistent with existing mailcrypt stuff would take more attention.) Thanks for any responses! Ken Manheimer ken...@gm... |
From: Fredrik N. <fc...@no...> - 2005-07-05 20:01:43
|
Ken Manheimer <ken...@gm...> writes: > In fact, for personal, day-to-day uses like journal entries and > such, i see symmetric encryption ... as being the default mode - > key-pair being somewhat heavyweight, ideally having a difficult > passphrase and choice restricted to only established keys. Symmetric encryption may be best for this sort of thing: it's not that public key is "better," but serves a different purpose. It depends upon whether you expect to have the same person encrypting and decrypting the data, or you expect others to decrypt what one has encrypted. For the first you just need a passphrase, for the second you need key distribution. From the sound of your application, symmetric would probably be the way to go, I'd think. > - whether i've missed something, and i can do symmetric-key > pgp/gpg encryption with mailcrypt without major contortions Alas, I've not gotten around to analyzing mailcrypt myself, so can't answer this. > - if not, whether it's so hard on purpose, and if so, why. I wouldn't think it's made hard on purpose, just it didn't fit the original intent (sending and receiving email) too well. > - If it is hard, but not on purpose, would any of you be willing to work > with me to make it easier to do symmetric-key encryption? (the > machinery would be pretty trivial, but keeping consistent with > existing mailcrypt stuff would take more attention.) I've no time for such a project right now, sorry! /Fredrik +----------------------------------------------------------------+ | Symeon | Fredrik Noon, Senior Software Engineer | | fc...@no... | Hifn, Inc. www.hifn.com | | www.noon.org | fn...@hi... +1 408 399 3630 | |-------------------+--------------------------------------------| | pgp key: <http://noon.org/keys/pgpkey.txt> 7840AC55 | +----------------------------------------------------------------+ =2D----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Ken Manheimer <ken...@gm...> writes: > In fact, for personal, day-to-day uses like journal entries and > such, i see symmetric encryption ... as being the default mode - > key-pair being somewhat heavyweight, ideally having a difficult > passphrase and choice restricted to only established keys. Symmetric encryption may be best for this sort of thing: it's not that public key is "better," but serves a different purpose. It depends upon whether you expect to have the same person encrypting and decrypting the data, or you expect others to decrypt what one has encrypted. For the first you just need a passphrase, for the second you need key distribution. From the sound of your application, symmetric would probably be the way to go, I'd think. > - whether i've missed something, and i can do symmetric-key > pgp/gpg encryption with mailcrypt without major contortions Alas, I've not gotten around to analyzing mailcrypt myself, so can't answer this. > - if not, whether it's so hard on purpose, and if so, why. I wouldn't think it's made hard on purpose, just it didn't fit the original intent (sending and receiving email) too well. > - If it is hard, but not on purpose, would any of you be willing to work > with me to make it easier to do symmetric-key encryption? (the > machinery would be pretty trivial, but keeping consistent with > existing mailcrypt stuff would take more attention.) I've no time for such a project right now, sorry! /Fredrik +----------------------------------------------------------------+ | Symeon | Fredrik Noon, Senior Software Engineer | | fc...@no... | Hifn, Inc. www.hifn.com | | www.noon.org | fn...@hi... +1 408 399 3630 | |-------------------+--------------------------------------------| | pgp key: <http://noon.org/keys/pgpkey.txt> 7840AC55 | +----------------------------------------------------------------+ =2D----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.4 (Cygwin) iD8DBQFCyuchAi4MWHhArFURAh3NAKC5urvcAPdsAQXUcJyMhRdRDqzHCACgpMyQ M/357gQYMtDVsEWSJt7D1eg=3D =3DqVM7 =2D----END PGP SIGNATURE----- |