I don’t remember the exact extent or know if it’s changed, but Radicle’s web interface definitely used javascript excessively at the time I posted the original proposal.

It sounds like I was confused about Typescript, I thought it was more closely related to js

Awesome, I think you’re the first person to give Cradicle a bit of a code audit. Thank you. I’ll check your feedback with a more C-focused programming community to see what others think before passing it onto the dev.

This reminds me of another reason I thought the dev made an interesting choice with C: by using C, we might attract “red team” volunteers to provide scrutiny.

• Guide: you can’t compete with github using a CLI guide while there’s no one click solution. With BitTorrent in its constant state of DDoS and fragmentation despite having easier seeding, radicle’s harder seeding makes the math even worse.
• Fork: it was up to radicle whether to implement the proposal in their mainline release or require a fork, but I don’t see how or why you’d pretend a fork wasn’t required after a fork is already the thing that happened faster (and a massively rewritten one, too).
• C rewrite: indeed, as I said to begin with that was an expansion by the dev that took the proposal, beyond the scope of the original requirements.

That isn’t really any more secure.

It is when javascript vulnerabilities are exploited, that’s why Tor Browser’s “safest” preset doesn’t allow js

• Tor: Our goal is to make seeding a git repo as simple as seeding a torrent, and Tor seems like it can indeed allow users to do it that straightforwardly with the bonus of not exposing everyone’s IP addresses to each other.
• Typescript: Cradicle’s web interface improves security by keeping all the complex functionality in the HTML server and serving pure scriptless HTML to web browsers instead.
• C: I’m not a coder but I thought it was pretty cool when the main dev Andrew K announced they’re using C, because I hear it allows more control over compiler behavior, and in the long run of course there should be radicle apps in multiple coding languages, like there are BitTorrent apps in multiple coding languages.

Edit - I’ll leave my original reply crossed out, but it doesn’t feel like a good reply to me and I think sometimes I don’t know how to properly respond to people these days.

Stop with the gish gallops.

• I should not need to address this many points because you keep pulling random nonsense out of nowhere. Look how fucking long this list of bullet points is just for one reply. Do some amount of double-checking your own thoughts before making me think through every single thought that pops into your head.
• You’re incorrect on your first point. Non-relay “peers” don’t stop Tor from scaling.
• When you say “from the start” it seems to imply I started “arguing for anonymity” at some point. I didn’t. This is an electronic network.
• Instead of explaining how hyphanet is supposed to be better than Tor for p2p git repos, you kept dwelling on conflating “privacy” with something users can get from a widespread software network on today’s hardware.
• Tor being “slow” is “intentional” for the same reason, I don’t get what makes you think that’s groundbreaking.
• What’s not the place to delve into what?
• This pattern keeps showing up: the part about BitTorrent seems like you’re conflating “anonymity” with something users can get out of a network like this, and not actually getting to whatever your point is about BitTorrent.

If Tor gets more users, they can run Tor nodes. The network scales to the traffic because it’s decentralized. You’re making up a problem that isn’t there. The Tor project does not recommend against creating traffic using onion services, again it wouldn’t make sense, there would be nothing to do with onion services then.

I’m really not understanding your point about anonymity. Why would we want every user in a P2P network to expose their clearnet IP address to every other user? It seems like if you flipped the release dates of Tor and BitTorrent, then Bram Cohen probably would have designed BitTorrent with Tor in mind from the start and it wouldn’t burden exit nodes so much. People would only use it without Tor when speed is more of a priority than security. But instead BitTorrent came first and already relied on an out-of-Tor network.

You seem to be suggesting 3 things:

• I thought onion services got gigabit speeds, for some reason
• It’s my fault for some reason if people with gigabit client/server connections available insist on using Tor and then complain about speed
• Aside from the mostly-unrelated traffic issue I was reminded of - “torrents generally strain exit nodes” like Tor officially says - there’s also a more pertinent issue of “onion services should never be used for traffic because that would strain Tor”

Are we on the same page with these 3 bullet points, or am I still not understanding you?

I don’t get what you mean by “traffic-pressured.” If you’re referring to the Tor devs recommending against the use of torrents, that’s because torrents typically involve out-of-Tor traffic which burdens exit nodes. Cradicle isn’t torrents, the Tor devs don’t recommend against using onion addresses for their intended purpose of traffic within the Tor network, that wouldn’t make much sense.

Regardless, if you have a better way to implement one-click seeding like bittorrent clients have, feel free to suggest it or build it.

That doesn’t sound right. Maybe design/implementation details would make it less nonsensical.

The proposal used Tor’s onion addresses as an answer to Radicle (unlike BitTorrent) requiring seeders to have fixed IP addresses or domain names (edit - or preconfigured onion proxies or something making onion addresses complex to use). Another reply suggests this problem with Radicle has been fixed or they’ve started fixing it while this project was in development. I haven’t really kept up with the news on Radicle in that timeframe.

sounds like a bad joke

I don’t see why. If radicle is a serious decentralized project at the scale of potentially replacing github, it should of course end up with multiple implementations in multiple programming languages. Can you clarify your view more?

At the time I made the proposal, I couldn’t figure out a simple way to seed repos with Radicle like I was familiar with from GUI-based BitTorrent clients, and when I asked people in the community why it was so complicated, the answer I ended up with was that it was due to domain names / static IP addresses being required instead of onion addresses (or other decentralized cryptographic addresses)

Has that changed since then? Doesn’t quite seem clear to me from the link you posted

Edit - after thinking about it a little more, I remember it may have been possible to use Tor via a proxy at the time I made the proposal, but it didn’t work fully and wasn’t anywhere near a 1 click solution like bittorrent. Still not sure if that’s changed

Nah, anyone that pretends BitTorrent is worse than their server-based thing is just foolish

You are correct

Some might be more trustworthy than others though. I think Eden seems like a pretty good project run by good devs