Announcing ClearCrypt: a new transport encryption library

mail.mozilla.org

27 points·dbaupp··10 comments

What's the point of announcing something that doesn't exist?

Don't waste your time on the link. It's pure vaporware.

Actually vaporware usually usually has a design, but this is more like "write something in rust, but we're not sure yet what (but it sounds good)".

So let's call it "nothingware"

I suppose to gather crypto people to contributes via patches, reviews or whatever else.

We're aiming for a project where all commits receive public scrutiny. This includes how the project is governed.

We're discussing contributing guidelines here:

https://github.com/clearcrypt/clearcrypt/pull/2

Initial code PR with a ChaCha20 implementation here:

https://github.com/clearcrypt/clearcrypt/pull/3

It would be nice if some of the items in the pre-commit checklist from [1] could be checked at compile time somehow.

[1] https://github.com/clearcrypt/clearcrypt/pull/3#issue-327877...

CurveCP sounds good, but what's Noise? And could CurveCP be improved by using QUIC instead of UDP?

https://github.com/trevp/noise/wiki

QUIC already has a lot of TLS support baked in afaik...

NaCL already exists and CurveCP is a part of that, why not just add a Rust port? (and a protocol for certificate transport)

> Emphasis will be placed on simplicity, clarity, and audibility.

Most projects start with this as an implicit goal. Unfortunately they tend to grow out of control as the code base gets larger.

We aim to make it an explicit goal, rather than implicit. Wherein other projects this is a "would be nice", in this project it's an absolute necessity for the project's security and therefore success.

Wow, you can announce stuff before it's actually built now?

Reply on news.ycombinator.com