4 ms·

Oops, sorry, mixed up the titles. I meant to write PC where I wrote AC. If PC disavows AC, then I've failed to find the disavowal by looking up the pointers to

by gjm11 18y ago

Oops, sorry, mixed up the titles. I meant to write PC where I wrote AC. If PC disavows AC, then I've failed to find the disavowal by looking up the pointers to AC in the index of PC. There's something nearer to a disavowal in "Secrets and Lies", but it still doesn't go further than (I paraphrase) "I focused on the algorithms, but actually other stuff matters more". (Both books mention that plenty of very bad systems have been built by people who read AC and thought they were therefore experts, but I don't think it's fair to blame that on AC.) There's a reason why I put "arguably" in front of "definitive" :-). But the point is that what got Schneier famous was writing a big fat book, with lots of technical content, that a lot of people read and were impressed by. That may be less solid than writing a big book that deserves to impress everyone, but it's not at all the same thing as pure blogging bloviation. (AC doesn't seem so very bad to me, aside from being out of date and being too much of an unassimilated algorithm-dump, but then I'm a generalist rather than a security professional.)

gjm11 18y ago
(For the avoidance of doubt, not that it particularly matters, the book I meant to call "arguably definitive" was Applied Cryptography, not Practical Cryptography. I agree that no one would think PC is definitive even though it's a wiser book than AC, simply because it's so much more selective.)
tptacek 18y ago
You're right; I've overstated (I'm not making up that he's essentially disavowed it, but I can't find the cite I'm looking for, so assume it's just hyperbole). Here's the passage I was thinking of in "Practical": Among cryptographers, Bruce's first book, Applied Cryptigraphy, is both famous and notorious. It is famous for bringing cryptography to the attention of thousands of people. It is infamous for the systems that these people then designed and implemented on their own. The problems with AC include: * No attention given to any of the practical vulnerabilities in cryptosystems, so that you could deploy code directly from the book and still have it be vulnerable to ECB cut-and-paste or parameter tampering. * A candy shop of random ciphers without any context as to why one would be chosen over the other, with varying degrees of detail provided for each. * Descriptions of protocols that are largely obsolete or discredited, without warnings or disclaimers or, really, any actual didactic purpose. "Applied" seemed great to me too, but then I became a practioner (though by no means an expert). Even "Practical" lacks detail on a lot of major crypto issues --- side-channel attacks, parameter tampering, the safe use of public key primitives and signature validation --- that actually occur in real systems.
- kaens 18y ago
  I've read AC, and found it good for getting a handle on what's going on in cryptosystems. It definitely didn't leave me feeling like I was anywhere near prepared to implement one. (And the book's a bit dated). Since you seem to have a lot of experience in this area, what would you recommend as really good reading?
  - tptacek 18y ago
    You want to ask Colin Percival (hn://cperciva) this question, since Colin's a professional cryptographer and I'm a software security person. That said, I like "Practical Cryptography" (not Applied Cryptography) and I don't know of a good book that covers side channel attacks or parameter tampering; you just have to be very interested in the topic, follow the news, and start attacking systems. I would say one of the best things about "Practical" is that it will leave you feeling even less prepared to implement a cryptosystem on your own, even though you'll know much more about how to do it.
    - kaens 18y ago
      Thanks. I'll give PC a read, and perhaps drop cperciva a line. I'm pretty fascinated by cryptography, and I like to be decently well-informed, but I don't see myself focusing on it anytime soon.

by gjm11 18y ago

gjm11 18y ago

(For the avoidance of doubt, not that it particularly matters, the book I meant to call "arguably definitive" was Applied Cryptography, not Practical Cryptography. I agree that no one would think PC is definitive even though it's a wiser book than AC, simply because it's so much more selective.)

tptacek 18y ago

You're right; I've overstated (I'm not making up that he's essentially disavowed it, but I can't find the cite I'm looking for, so assume it's just hyperbole). Here's the passage I was thinking of in "Practical": Among cryptographers, Bruce's first book, Applied Cryptigraphy, is both famous and notorious. It is famous for bringing cryptography to the attention of thousands of people. It is infamous for the systems that these people then designed and implemented on their own. The problems with AC include: * No attention given to any of the practical vulnerabilities in cryptosystems, so that you could deploy code directly from the book and still have it be vulnerable to ECB cut-and-paste or parameter tampering. * A candy shop of random ciphers without any context as to why one would be chosen over the other, with varying degrees of detail provided for each. * Descriptions of protocols that are largely obsolete or discredited, without warnings or disclaimers or, really, any actual didactic purpose. "Applied" seemed great to me too, but then I became a practioner (though by no means an expert). Even "Practical" lacks detail on a lot of major crypto issues --- side-channel attacks, parameter tampering, the safe use of public key primitives and signature validation --- that actually occur in real systems.

kaens 18y ago
I've read AC, and found it good for getting a handle on what's going on in cryptosystems. It definitely didn't leave me feeling like I was anywhere near prepared to implement one. (And the book's a bit dated). Since you seem to have a lot of experience in this area, what would you recommend as really good reading?
- tptacek 18y ago
  You want to ask Colin Percival (hn://cperciva) this question, since Colin's a professional cryptographer and I'm a software security person. That said, I like "Practical Cryptography" (not Applied Cryptography) and I don't know of a good book that covers side channel attacks or parameter tampering; you just have to be very interested in the topic, follow the news, and start attacking systems. I would say one of the best things about "Practical" is that it will leave you feeling even less prepared to implement a cryptosystem on your own, even though you'll know much more about how to do it.
  - kaens 18y ago
    Thanks. I'll give PC a read, and perhaps drop cperciva a line. I'm pretty fascinated by cryptography, and I like to be decently well-informed, but I don't see myself focusing on it anytime soon.