#tlsnotary-chat

http://www.wired.com/2015/11/zerocoin-startup-r... i don't like the “Bitcoin is HTTP for money. We are HTTPS,” part

i don't like the part where zooko thinks the best thing to do is to start another shitcoin :)

well yes, but understandable

waxwing: where did you guys get the 'irc.py' or you just wrote it from scratch?

it was belcher's own i believe

i did suggest looking into an existing lib, but when i read his code, he seemed to know what he was doing

oh wow

random_nick is...precious

cute

oh that was later

nothing particularly interesting there imo

sorry, you are talking about JM, right?

yeah you must be, fine, no misunderstanding

yes

so the message/unsigned tx parsing is done in irc.py as well

well transactions and signatures aren't *parsed* there

okay, we can say it's just message parsing

the fact that there is message decryption calls in there is a pain point which we discussed here a couple of days ago (and indeed discussed it many months ago)

mildly interesting: https://plus.maths.org/content/ramanujan

speak of that, i wanted to read Gauss's disquistiones arithmeticae a few days ago, and was surprised to find that the English translation is still copyrighted, while the Latin original is in public domain :)

what's in there that piqued your interest i wonder?

does it have his FTA proof?

well, much like why people still read Landau and Lifshitz+can't be bothered to look up what number theory textbook is good

so to save time, you'll just learn Latin :)

otoh, i think people would read Gauss's diary out of genuine mathematical interest

waxwing: quite the opposite, it's always nice to find some however trivial RL usages of Latin :)

ah, Putney

it's a nice delight to know that there are mathematicians taking time to peruse past mathematicians's manuscripts

hmmm, okay, so we x^3+y^3=z^3 a classical elliptic curve is not far away

but what will you come up with when n>3?

bertani, i think ethereum must have an opcode similar to CHECKSIG in bitcoin, so you can have your oracle produce an ethereum-compatible EC signature and use that opcode to verify it.

waxwing, proslogion, I am beginning to lean towards that containerizing the notary (be it docker or lxc) is an overkill. There is enough security as it is if notary and sigserver are run under two diff user accounts, thus if notary is exploited, he cannot read the sigserver's privkey. With this setup the attacker would have to A) exploit a bug in our py code and B) exploit a bug python executable and after C) ex

ploit a 0-day kernel vuln in order to get the permissions to read the privkey. Adding a container around the notary would add a 4th layer of protection. I dont believe it's worth the effort tbh.

dansmith_btc: well, no expert, but i agree; i do like the second server for signing key idea, but i don't consider it necessary to have a working implementation.