WelcomeUser Guide
ToSPrivacyCanary
DonateBugsLicense

©2026 Poal.co

838

See here for a description of what I'm doing: https://poal.co/s/technology/569412

Today, I'm fasting so I'm making an earlier log entry before I lose all my mental energy.

I've been looking at p2p lately. I thought about it, and using a centralized p2p forum to power the thing seems to be going down a rabbithole. First because of what I've seen the stuff that's out there is very bloated, poorly documented, and hard to customize, and since I'm a beginner it makes it even more so. Besides, even if I were to succeed, then so what? All I'd get out of it is a good prototype, but I'd still have to move to somethig more decentralized.

So, I'm shifting gears and going p2p again. Maybe it's obvious, but I realized when trying to understand what a library does, it's best to look at its dependencies first, so if I want to understand hypercore, I look and see it depends on hyperstream and secret-stream, and then look to see that hyperstream depends on hyperstream/dht, etc. Then I just look at dht and secretstream and then go upwards from there.

So, from what I found, there is no way to create a p2p network without any fixed addresses. Even magnet links have bootstrapping nodes hardcoded in the client. So what I plan to do is to piggyback on bittorrent's magnet protocol. That way, in order to knock out captain dirgo through DHT, you'd have to knock out the whole bittorrent infrastructure.

The big elephant in the room then is spam. There are a couple of projects I want to look at that are also based on hypercore that would have to deal with spam, that I want to look at to get ideas. One is https://github.com/Telios-org which is a p2p email service, and the other is https://cabal.chat/ which is a p2p chat network. I found these here: https://dat-ecosystem.org/

My own idea is to use something like hashcash, so that to post a message, you'd have to mine some data akin to a bitcoin miner. This would be as light as possible, for a url with zero comments, or few views, maybe mining wouldn't be needed at all, but one with a lot of comments, you would have to mine for awhile. Also along with that, having trust networks, so that if your user is trusted by another node on the network it could post without mining. That other node would have rules, such as having to solve a captcha to initialize the node, and other rules like not posting too quickly.

In the end, its up to each individual user about what messages they want to see or not, so when using the extension you could personally dial in how much spam you are willing to tolerate.

See here for a description of what I'm doing: https://poal.co/s/technology/569412 Today, I'm fasting so I'm making an earlier log entry before I lose all my mental energy. I've been looking at p2p lately. I thought about it, and using a centralized p2p forum to power the thing seems to be going down a rabbithole. First because of what I've seen the stuff that's out there is very bloated, poorly documented, and hard to customize, and since I'm a beginner it makes it even more so. Besides, even if I were to succeed, then so what? All I'd get out of it is a good prototype, but I'd still have to move to somethig more decentralized. So, I'm shifting gears and going p2p again. Maybe it's obvious, but I realized when trying to understand what a library does, it's best to look at its dependencies first, so if I want to understand hypercore, I look and see it depends on hyperstream and secret-stream, and then look to see that hyperstream depends on hyperstream/dht, etc. Then I just look at dht and secretstream and then go upwards from there. So, from what I found, there is no way to create a p2p network without any fixed addresses. Even magnet links have bootstrapping nodes hardcoded in the client. So what I plan to do is to piggyback on bittorrent's magnet protocol. That way, in order to knock out captain dirgo through DHT, you'd have to knock out the whole bittorrent infrastructure. The big elephant in the room then is spam. There are a couple of projects I want to look at that are also based on hypercore that would have to deal with spam, that I want to look at to get ideas. One is https://github.com/Telios-org which is a p2p email service, and the other is https://cabal.chat/ which is a p2p chat network. I found these here: https://dat-ecosystem.org/ My own idea is to use something like hashcash, so that to post a message, you'd have to mine some data akin to a bitcoin miner. This would be as light as possible, for a url with zero comments, or few views, maybe mining wouldn't be needed at all, but one with a lot of comments, you would have to mine for awhile. Also along with that, having trust networks, so that if your user is trusted by another node on the network it could post without mining. That other node would have rules, such as having to solve a captcha to initialize the node, and other rules like not posting too quickly. In the end, its up to each individual user about what messages they want to see or not, so when using the extension you could personally dial in how much spam you are willing to tolerate.

(post is archived)

[–] 0 pt

Looks like you've turned /s/Programming into the Captain Dirgo sub now. How about you make a Captain Dirgo sub and post in there so people who are interested in it can subscribe and /s/Programming can return to being mostly ignored?

[–] 0 pt (edited )

I'd like to do what you suggest, but I am only level 6 and I need to be level 10 to make a captaindirgo sub. If this is annoying people I'll stop.

Edit: Talked to AOU and he helped me out.