FRESH

Hacker News

Home

Understanding ECDSA

96 points by avidthinker

by joezydeco

1 subcomments

Your definition of "basic math" greatly differs from mine...

by mkornaukhov

1 subcomments

I've recently understood how RSA works and thought it was a cool achievement. But this article with "basic" math... Not so enjoyable for just a dev =)

by oleganza

3 subcomments

ECDSA is a horrible workaround for patent on Schnorr signatures. Here's my talk from 2019 about the issue.
https://www.youtube.com/live/2IpZWSWUIVE?si=-LRRbU2mJgL9LiNP...

by atoav

1 subcomments

A big part of what makes maths hard for non-math-people is math notation. That is like if a well versed python programmer told you it is as simple as:
```
  result = [[{"x": x, "y": y, "v": (x+y if (x+y)%2==0 else None)} for (x, y) in row] for row in data]  
  
```
Now with a bit of reasoning another programmer that hasn't used python might be able to figure out what that means. But what if my audience is non-programmers? The moment they encounter the first unexplained square brackets and then an opening curly brace it will essentially feel like telling them: "Here is a riddle for you" or potentially even like "I expect you to know this, dummy".
Not that this text was particular bad in that regard, but I wish more math people had a heightened awareness of the fact that for many the hard part is not understanding the concept (e.g. fourier transformation), but the foreign looking signs mathematicians have decided to use to write them down.
That is as if someone explaines the way to the next train station to you in a foreign language. The hard part isn't understanding the way, it is understanding the noises that are supposed to make up the description.
And as a programmer who from time to time has to translate maths into discrete programs (or the other way around) the hard part was always parsing the notation and when I figured it out I was usually like: "Ohh, this is just a simple algorithm doing that.
So if you want to explain a math concept to programmers you should chose one of two routes:
(A) Stay with your notation and explain every character that isn't visible on a regular keyboard in length and gently lead the reader into being able to read the notation or
(B) let go of the notation and first explain what it does and how, e.g. for our FFT example: FFT slices your list of values into frequency buckets, figures out how much of each frequency is present, and returns those strengths as numbers. And then you can work backwards from that understanding towards your notation explaining which sign relates to which part of the concept (e.g. to the number of buckets).
I would prefer the latter, since it explains both the concept and gives the mathematician a chance to explain how and why math notation can be useful on top, e.g. to figure out certain properties of the method that may even have practical implications.

by NetMageSCW

1 subcomments

The article would be a lot better if it was what it said on the tin, instead of being filled with lots of unnecessary (as described in the article) digressions. If you couldn’t restrain yourself to sticking to the subject, at least put the digressions behind links or footnotes or pop-ups where they don’t detract from reading about the actual claimed intended subject.

by some_furry

1 subcomments

In addition to the malleability attack (high-S and low-S both being valid for a given value of R), ECDSA doesn't provide a property called exclusive ownership: https://soatok.blog/2023/04/03/asymmetric-cryptographic-comm...
In contrast, EdDSA (which is based on Schorr signatures) does, by construction: the public key is included in one of the hashes, which binds the signature to a particular public key.
I haven't investigated whether cryptocurrency's use of Schnorr satisfies this property or not. (Indeed, I do not care about cryptocurrency at all.) So it's an exercise to the reader if it's satisfactory or not :3

0 subcomment