[Monotone-devel] Re: fatal: Botan::PRNG_Unseeded

[Lapo Luchini]

Lapo Luchini wrote:
> ECDSA signatures would be so small

Which is of course of PARAMOUNT IMPORTANCE, because signatures right now 
use almost 4% of database space. That's not acceptable.
(yes, I'm kidding)

% mtn db exec 'SELECT SUM(LENGTH(signature)) FROM revision_certs;'
5861887
% bc
5861888/148422656
.03949456341759576112

> ECDSA signatures would be [...] so fast

...and even this is strictly not true at the current 80 bit security 
level we're using (1024 bit RSA is around 80 bit of security and 3.2x 
faster than ECDSA-256p), though it would be actually faster at higher 
security levels (e.g. ECDSA-256p has 128 bit of security and is 2.4x 
faster than the equivalent RSA 3072 bits; and produces signatures 512 
bit long, vs 3072 bit).

    Lapo