My attempt to explain Christian's Torrent-based idea
Let me explain what I take to be the value proposition of Christian’s proposal. I was sort of waiting for him to do this 🙂 but I can do it very briefly just to get the idea out there.
A public-private key pair enables content that is published with a public key to be cryptographically proven to originate only with a person who has his hands on the private key, and this can be proven without giving anyone else access to the private key.
Lord knows I’m not a cryptographer, but I think I understand this much: if you have a private key (only you have it) and a piece of content that you want to claim is yours, you can use a “hashing algorithm“ to output a “Hash“ and this becomes useful to others, because they can take your public key, and use it to show that your content was hashed using the private key. (Have I got it right?)
This is relevant to identity, because a public key and private key make up a pair. So this becomes important, says Christian, for identity and this project in the following way. With the same public-private key pair, a person can make various claims about his identity, in some standard format. He might be lying, but at least we will know if there is an interesting person who is lying because the public key will be associated with different social media accounts and what not.
And we can include that identity data alongside content which is hashed as well, and then this can be passed around in torrent files which zip up everything: public key, identity, hash, and content files. We use torrent files because the torrent network enables us to distribute the content via participant computers and servers in a P2P way, which makes the content as uncensorable as content on the rest of the Torrent networks.
So, so far, I’m liking this idea quite a bit. It is 100% compatible with my original vision, it just makes massive technical improvements that I hadn’t anticipated but knew needed to be made anyway.
Christian tells me this is a very simple and elegant solution, which makes me suspicious. 🙂
I would have one question: can’t IPFS do the same thing as the Torrent network, and might it be better?
I thought we were just going to have to have aggregators like blog aggregators, which have a separate API function, and ID established through domain names or something. But basically, letting people publish in Tor form with public keys and hashes allows content to be aggregated and distributed and verified as coming from their publicly declared identities.
This becomes very handy for organizations like the AMA which can provide lists of public keys and identity attributes they associate with those public keys (again, like names, degrees, certifications, etc.).
The above was first written in the Slack #dev channel. Ram's remarks there were:
@Larry Sanger You’ve got it right regarding public key cryptography.
It will be the fundamental component of self-sovereign identity as well as content author verification.
What’s interesting is that if at some higher level commenting on articles was made possible, and people could comment using their self-sovereign identities then their comments could also count as published “expertise” in a way.
Again, for now, I’m ignoring how ratings would work :slightly_smiling_face:
One thing to note regarding torrents or IPFS is that guaranteed content availability requires continuous hosting. So e.g. if we’re using torrents then every bit of content needs to have atleast 1 node seeding it in order to guarantee availability. Keeping in this mind, it’s quite likely that initially the KSF will be running 1 or more nodes to ensure access. But anyone else would be able to run a node and connect to the KSF nodes (thus building out the network) and eventually hopefully other large orgs will get involved and share the load, so to speak.
I think the goal of IPFS is different to that of Torrents but from a technical standpoint they are achieving something very similar. IPFS is about global, uncensorable storage whereas Torrent tech was initially developed to enable sharing content as quickly and as bandwidth-efficiently as possible with a large audience.
But I’d be interested in @Christian Gribneau’s thoughts on this!
And also Hampson's remarks:
what if there was a lite box that had personal use like web server / storage but also functioned as a node
it’s not incredibly simple because it should have mass adoption backdoor basically for that - something like using RSS/Atom would, or .torrent infrastructure vs. IPFS from what i understand
$TRX and $BTT are strongly supportive of torrent tech for this reason and why they’re easily going to grow to top 10 if not already, Bittorrent org got behind it to basically incentivize active seeding across the .torrent network
i think that’s a really curious and interesting way to develop this project but never considered it tbh