Re: LibTomCrypt

mm wrote:
Tom St Denis a écrit :
Pascal is an inferior [hehehehe] language. I imagine you could write a

Sorry, I nearly missed this one :-)

Seriously dropping the ball man...

Consider the following code (It was in your file "ecc.c" and I have to
thank you for having suppressed it from your lib since, now, I can talk
about it):

I supressed it because I didn't want Certicom suing me.

Assuming p is prime, p = 3 mod 4 and pt is a valid point on the curve,
how do you explain that *all* the previous C code can be replaced by a
*single* Pascal (this inferior language) instruction?

Yeah that's nice and all. Except not all of the curves are 3 mod 4.

Result := (1 - Jacobi(pt.y,p)) shr 1;

I didn't have a Jacobi function at the time IIRC. I also don't see
what that does.

SQRTs are a^(p+1)/4 when p is 3 mod 4. Jacobi is congruent to
a^(p-1)/2 which is not the same thing.

All Jacobi will do is tell us whether the y in

y^2 = x^3 - 3x + b

is a QR or not. What we want to know is which root of (x^3 - 3x + b)
is y.

Personally I can see only one explanation: what makes a program is good
or not is much less the language used than the way it is used.

Maybe you should explain your math a bit more.