Quoting Andrew McGlashan (andrew.mcglashan@affinityvision.com.au):
Now known as just Keyring for Palm OS, they've dropped GNU in the name, but the SF link still needs it; it would be good if it redirected to: http://keyring.sourceforge.net - oh well.
I _did_ mention that its name is Keyring. Yes, it did get renamed, early on. It's not all that interesting a story, which is why I skipped that bit.
I don't trust 3DES and 112 bits doesn't seem enough for my liking, that's only really 2DES anyway.
All standard 3DES implementations have always been effetively 112 bits, because of the chaining of three DES keys and a meet-in-the-middle problem. And yet that's been very robust.
It also looks like your password is crackable against the MD5 hash.... where the 32 bit random salt comes in seems quite unclear to me from the description.
Um, I might be missing what you're saying here, but this doesn't sound right. md5 as a hashing algorithm is showing some weaknesses, relatively speaking, and it certainly wouldn't be the right choice for a new codebase, which is why the two coders say they're going to use HMAC-SHA1 for the next version. _However_, even disrecommended for new implementations as it is, md5 is in no way easily 'crackable', and even finding collisions at a higher rate than initially hoped for (its main known problem area) isn't going to be very useful in brute-forcing a Keyring DB file. Note that the 2.0-ore6 beta is available and highly usaable. That's the one that switches to 168-bit 3DES (up from 112) and HMAC-SHA1 hasing iteratively applied using PBKDF2 (RFC 2898) key generation. Getting back to 3DES: People in open source are often surprised to hear that newer cryptographic algoirthms are (as a general rule) less trusted than older ones, in cases where both have been holding up pretty well to expert tinkering. That's because the best metric available is how long each has withstood determined expert attacks. 3DES with keyring option 1 has been quite robust. NIST claims it's estimated to be praatical and reliable in production through around 2050, at this point, and I don't recall any major cryptographer disagreeing.
KeePass and TrueCrypt are what I use (for now), both seem much more secure than Keyring for Palm OS, albeit I don't use it with an airgapped device.
Yes, that's the problem I have with such software (despite good implementation): They're exposing the crown jewels on a workstation, where any user-level security breach could lose you everything. My approach inherently avoids that entire problem.