4 ms·

Asynchronous Group Messaging with Strong Security Guarantees [pdf]

56 points by lainon 9y ago

Sephr 9y ago
I'm so glad this is out. Myself and a few friends tried to tackle this same problem as part of an open source project years ago but we eventually shelved the project due to lack of time. This will be very useful if it gets integrated into Tox and other chat platforms.
cvwright 9y ago
A useful trick: Omit the .pdf from ePrint URLs to bypass the PDF and get the summary page with authors, abstract, bibtex, etc. https://eprint.iacr.org/2017/666 https://eprint.iacr.org/2017/666
new2424141 9y ago
666.pdf hmmmmmmm
- dfcab 9y ago
  Noticed same thing
kodablah 9y ago
> We implemented the ART protocol described in Section 5 in Java as a proof of concept. Anyone have a link to the repo for this?
- lucb1e 9y ago
  This is why I hate papers and love blog posts: they're less formal, to the point, and you bet that if they claim there's code, there will be a link. That said, looking for "Asynchronous Ratcheting Tree" in Google, I find this paper and a bunch of irrelevant stuff. You'd think it'd be in the readme of the Github project if there was one. Looks like it's just academic.
  - Ar-Curunir 9y ago
    Lol good luck finding rigorous and novel cryptography in blog posts. Different media are suited for different purposes.
    - toomim 9y ago
      Pdfs are suited for printing. That has nothing to do with rigor. You're conflating fashion with intellect. Example: https://people.xiph.org/~greg/confidential_values.txt https://people.xiph.org/~greg/confidential_values.txt
      - Ar-Curunir 9y ago
        Even that fairly technical document contains no formal definitions or proofs. It would also be much improved with actual mathematical notation. Again, cryptography is a subtle science. It needs proper setup, proper explanation and proper contextualisation. That's not what blogs are good for. Sure, write a blog to accompany a formal document like a paper, but don't replace one with the other. Finally, PDFs are only the most popular format for paper writing because LaTeX does not compile well to HTML, and LaTeX is the best language for mathematics.
        toomim 9y ago
        Not true. It does make formal definitions and proofs. Without being a PDF.
        Ar-Curunir 9y ago
        Where are these proofs? Where are the definitions? There's barely even pseudocode describing the algorithm. Honest question: have you read cryptography research papers before?
kamilner 9y ago
I'm one of the authors, happy to see this on HN. Just to let everyone know we're currently in the process of updating the paper and implementation, so there will be a more detailed version up in a couple weeks. There's also an independent implementation in Go here[0] though I don't know how complete it is. [0] https://github.com/bifurcation/treekeys https://github.com/bifurcation/treekeys

56 points by lainon 9y ago

I'm so glad this is out. Myself and a few friends tried to tackle this same problem as part of an open source project years ago but we eventually shelved the project due to lack of time. This will be very useful if it gets integrated into Tox and other chat platforms.

cvwright 9y ago

A useful trick: Omit the .pdf from ePrint URLs to bypass the PDF and get the summary page with authors, abstract, bibtex, etc. https://eprint.iacr.org/2017/666 https://eprint.iacr.org/2017/666

new2424141 9y ago

666.pdf hmmmmmmm

dfcab 9y ago
Noticed same thing

kodablah 9y ago

> We implemented the ART protocol described in Section 5 in Java as a proof of concept. Anyone have a link to the repo for this?

lucb1e 9y ago
This is why I hate papers and love blog posts: they're less formal, to the point, and you bet that if they claim there's code, there will be a link. That said, looking for "Asynchronous Ratcheting Tree" in Google, I find this paper and a bunch of irrelevant stuff. You'd think it'd be in the readme of the Github project if there was one. Looks like it's just academic.
- Ar-Curunir 9y ago
  Lol good luck finding rigorous and novel cryptography in blog posts. Different media are suited for different purposes.
  - toomim 9y ago
    Pdfs are suited for printing. That has nothing to do with rigor. You're conflating fashion with intellect. Example: https://people.xiph.org/~greg/confidential_values.txt https://people.xiph.org/~greg/confidential_values.txt
    - Ar-Curunir 9y ago
      Even that fairly technical document contains no formal definitions or proofs. It would also be much improved with actual mathematical notation. Again, cryptography is a subtle science. It needs proper setup, proper explanation and proper contextualisation. That's not what blogs are good for. Sure, write a blog to accompany a formal document like a paper, but don't replace one with the other. Finally, PDFs are only the most popular format for paper writing because LaTeX does not compile well to HTML, and LaTeX is the best language for mathematics.
      - toomim 9y ago
        Not true. It does make formal definitions and proofs. Without being a PDF.
        Ar-Curunir 9y ago
        Where are these proofs? Where are the definitions? There's barely even pseudocode describing the algorithm. Honest question: have you read cryptography research papers before?

kamilner 9y ago

I'm one of the authors, happy to see this on HN. Just to let everyone know we're currently in the process of updating the paper and implementation, so there will be a more detailed version up in a couple weeks. There's also an independent implementation in Go here[0] though I don't know how complete it is. [0] https://github.com/bifurcation/treekeys https://github.com/bifurcation/treekeys