Re: [OT] Firefox 3.0

Simian wrote:

Slower Than You wrote:

Again, FWIW, signcode uses MD5 or SHA1, both of which are public
algorithms, I have no idea what Wise Install uses (or indeed, what it
actually is), but I'd bet a fair amount that it's from the same set of
knowns aRC4/MD5/SHA1.

You are conveying your lack of experience. Sure, authenticode process
does use MD5 or, optionally SHA1 as a hashing algorithm but that's a
vanishingly small part of the process; a mere hash of an executable or
archive is hardly very useful on its own in this context is it? You
neglected to mention anything at all about the much more significant
PKC side of the equation, timestamping, and certification.

Wise Installer uses the CryptoAPI to generate it's authenticode
signatures, just like every other commercially-available code signing
utility I'm aware of (which is by no means all of them, but if someone
has written their own implementation it would probably be considered
"news"). Sure the algorithms used are well-known and published, but
only someone who has never tried it would consider the job of
implementing them securely something that "any old idiot can knock up".
It's worth noting that most attacks on cryptosystems are against the
*implementation* rather than the algorithms.

I have written my own implementations of RSA, ElGamal (both using my
own huge int library) MD5, ARC4 (everyone's written a CipherSabre -
it's a right of passage) and TwoFish over the years - my OCX wrapper
around the DLL implementation of TwoFish was downloadable via bruce
schnier's web site for a while (until, I admit, it was quickly replaced
with something with fewer bugs). Needless to say, I have therefore
communicated with industry leaders like him at various points in the
past. I don't doubt that you may have done some work in this field, but
so have I. God, I could email some source code straight from my dev
machine if you're having problems believing me.

Unlike you, I know enough to that downloading a huge int library and
"knocking up" a public key crypto implementation might not actually be
the wisest, or most trivial, thing to do.

Anyway, this is turning into a ***-size war now, and taking up a
disproportionate amount of time for little gain. So I'm not posting any
more.
--
Slower Than You
.