Comment by Uptrenda | Hacker News Reader

Comment by Uptrenda | Hacker News Reader

To me this just seems pointless because an implementation in Rust is going to have similar features to one in C, Java, Python, or any of the handful of languages where Tor already exists... I want to rep another project that I think would be more relevant to Tor's goals. It is unfunded yet should it exist: would have the potential to transform privacy on the web. The project is https://github.com/Ayms/node-Tor It's about implementing Tor in Javascript.

Why is this a good idea? Well, if you had a Javascript version of Tor you could quite literally import it on any website. It would be the easiest way to add privacy to millions of users sessions. It would be so ridiculously accessible and easy to use that it would probably lead to rapid innovation in the underlying tech. There is a lot of cool p2p tech already built into the browser that would be an amazing test bed.

Currently Google Chrome are also experimenting with a socket-like API that I believe would be a useful piece of the puzzle [for the project.] It just seems to me that marketing, networking, and being a tech normie is more important than actually having a good idea. So you see pointless projects like this get all the funding while projects like Ayms stay unknown and don't get any. Historical note: the hacker Aaron Swartz (tragically deceased - those who know... know. It was an awful waste) was involved in this project and suggested it as an idea. It's definitely a good one in my opinion.

The authors emphasize how the Rust compiler prevented them from introducing many potential bugs.

Sometimes it's better to have no security software and fully understand the implications than to have a security software with bugs unknown to you, which are tacitly exploited by your adversary.

This is a stunningly baffling idea given that any user that has half a clue is using Tor with JS disabled.

I see what you're saying. But there's no reason why you couldn't do the same thing with this design. You can build a basic DOM model that is allowed and only let that portion through. While the outer layer uses Javascript freely to run. You can decide what policy you want before displaying any output.

You could compile the rust version to WASM and run it in browser.

The issues preventing this from happening are actually tracked in the Arti repository.

That would have been an amazing project back in 2016 when WASM didn’t exist. Today a js implementation of Tor is pointless.

I don't think that Tor in javascript would improve anything. Tor in golang on the other hand would gain multithreading and ease of integration. It would drop in nicely with the Golang standard library.

A Go version wouldn't improve anything upon the Rust version either. At this scale, Go's simplicity would be a hindrance more than a help, development velocity would suffer compared to Rust. Integration would only be possible in the Go ecosystem, whereas Rust can expose a C API so that Arti can be used (eventually) as the engine for implementations in other languages (Go, Python, JS, etc). Multithreading and overall performance can be better optimized in Rust.

Tor is critical security software with a lot of implementation gotchas. Just like you shouldn't write your own crypto, you shouldn't write your own Tor. Arti is developed by contributors of the original C Tor, I wouldn't go near a Js/Go/whatever implementation from a team with less credentials.

Reply on news.ycombinator.com