Gemini is a new internet technology supporting an electronic library of interconnected text documents. That’s not a new idea, but it’s not old fashioned either. It’s timeless, and deserves tools which treat it as a first class concept, not a vestigial corner case. Gemini isn’t about innovation or disruption, it’s about providing some respite for those who feel the internet has been disrupted enough already. We’re not out to change the world or destroy other technologies. We are out to build a lightweight online space where documents are just documents, in the interests of every reader’s privacy, attention and bandwidth.
I’ve personally played with Gemini a few months ago, and now want a new Internet as opposed to a new Web.
Replace IP protocols with something better. With some kind of relative addressing, and delay-tolerant synchronization being preferred to real-time connections between two computers. So that there were no permanent global addresses at all, and no centralized DNS.
With the main “Web” over that being just replicated posts with tags hyperlinked by IDs, with IDs determined by content. Structured, like semantic web, so that a program could easily use such a post as directory of other posts or a source of text or retrieve binary content.
With user identities being a kind of post content, and post authorship being too a kind of post content or maybe tag content, cryptographically signed.
Except that would require to resolve post dependencies and retrieve them too with some depth limit, not just the post one currently opens, because, if it’d be like with bittorrent, half the hyperlinks in found posts would soon become dead, and also user identities would possibly soon become dead, making authorship check impossible.
And posts (suppose even sites of that flatweb) being found by tags, maybe by author tag, maybe by some “channel” tag, maybe by “name” tag, one can imagine plenty of things.
The main thing is to replace “clients connecting to a service” with “persons operating on messages replicated on the network”, with networked computers sharing data like echo or ripples on the water. In what would be the general application layer for such a system.
OK, this is very complex to do and probably stupid.
It’s also not exactly the same level as IP protocols, so this can work over the Internet, just like the Internet worked just fine, for some people, over packet radio and UUCP or FTN email gates and copper landlines. Just for the Internet to be the main layer in terms of which we find services, on the IP protocols, TCP, UDP, ICMP, all that, and various ones and DNS on application layer, - that I consider wrong, it’s too hierarchical. So it’s not a “replacement”.
IP is the most robust and best protocol humanity ever invented. No other protocol survived the test of time this well. How would you even go about replacing it with decentralization? Something needs to route the PC to the server
I don’t want client-server model. I want sharing model. Like with Briar.
The only kind of “servers” might be relays, like in NOSTR, or machines running 24/7 like Briar mailbox.
IP. How would I go about replacing it? I don’t know, I think Yggdrasil authors have written something about their routing model, but 1) it’s represented as ipv6, so IP, 2) it’s far over my head, 3) read the previous, I don’t really want to replace it as much as not to make it the main common layer.
client-server model. I want sharing model. Like with Briar
Guess what
Briar itself, and every pure P2P decentralized network where all nodes are identical… are built on Internet Sockets which inherently require one party (“server”) to start listening on a port, and another party (“client”) to start the conversation.
Briar uses TCP/IP, but it uses Tor routing, which is IMO a smart thing to do
Yep! That was exactly the protocol on my mind. One thing, though, is that the Fediverse would need to be ported to Gemini, or at least for a new protocol to be created for Gemini.
If it becomes popular enough that it’s used by a lot of people then the bots will move over there too.
They are after data, so they will go where it is.
One of the reasons that all of the bots are suddenly interested in this site is that everyone’s moving away from GitHub, suddenly there’s lots of appealing tasty data for them to gobble up.
It shouldn’t be too hard, and considering private key authentication, you could even use a single sign-in for multiple platforms/accounts, and use the public key as an identifier to link them across platforms. I know there’s already a couple proof-of-concept Gemini forums/BBSs out there already. Maybe they just need a popularity boost?
Like Gemini?
From official Website:
It’s not the most well thought-out, from a technical perspective, but it’s pretty damn cool. Gemini pods are a freakin’ rabbi hole.
Won’t the bots just adapt and move there too?
I’ve personally played with Gemini a few months ago, and now want a new Internet as opposed to a new Web.
Replace IP protocols with something better. With some kind of relative addressing, and delay-tolerant synchronization being preferred to real-time connections between two computers. So that there were no permanent global addresses at all, and no centralized DNS.
With the main “Web” over that being just replicated posts with tags hyperlinked by IDs, with IDs determined by content. Structured, like semantic web, so that a program could easily use such a post as directory of other posts or a source of text or retrieve binary content.
With user identities being a kind of post content, and post authorship being too a kind of post content or maybe tag content, cryptographically signed.
Except that would require to resolve post dependencies and retrieve them too with some depth limit, not just the post one currently opens, because, if it’d be like with bittorrent, half the hyperlinks in found posts would soon become dead, and also user identities would possibly soon become dead, making authorship check impossible.
And posts (suppose even sites of that flatweb) being found by tags, maybe by author tag, maybe by some “channel” tag, maybe by “name” tag, one can imagine plenty of things.
The main thing is to replace “clients connecting to a service” with “persons operating on messages replicated on the network”, with networked computers sharing data like echo or ripples on the water. In what would be the general application layer for such a system.
OK, this is very complex to do and probably stupid.
It’s also not exactly the same level as IP protocols, so this can work over the Internet, just like the Internet worked just fine, for some people, over packet radio and UUCP or FTN email gates and copper landlines. Just for the Internet to be the main layer in terms of which we find services, on the IP protocols, TCP, UDP, ICMP, all that, and various ones and DNS on application layer, - that I consider wrong, it’s too hierarchical. So it’s not a “replacement”.
IP is the most robust and best protocol humanity ever invented. No other protocol survived the test of time this well. How would you even go about replacing it with decentralization? Something needs to route the PC to the server
I don’t want client-server model. I want sharing model. Like with Briar.
The only kind of “servers” might be relays, like in NOSTR, or machines running 24/7 like Briar mailbox.
IP. How would I go about replacing it? I don’t know, I think Yggdrasil authors have written something about their routing model, but 1) it’s represented as ipv6, so IP, 2) it’s far over my head, 3) read the previous, I don’t really want to replace it as much as not to make it the main common layer.
Guess what
Briar itself, and every pure P2P decentralized network where all nodes are identical… are built on Internet Sockets which inherently require one party (“server”) to start listening on a port, and another party (“client”) to start the conversation.
Briar uses TCP/IP, but it uses Tor routing, which is IMO a smart thing to do
I’m talking about Briar used over BT.
Yep! That was exactly the protocol on my mind. One thing, though, is that the Fediverse would need to be ported to Gemini, or at least for a new protocol to be created for Gemini.
If it becomes popular enough that it’s used by a lot of people then the bots will move over there too.
They are after data, so they will go where it is.
One of the reasons that all of the bots are suddenly interested in this site is that everyone’s moving away from GitHub, suddenly there’s lots of appealing tasty data for them to gobble up.
This is how you get bots, Lana
Yes, I know. But, while trying to find a way to bomb the AI datacenters (/s, hopefully it doesn’t come to this), we can stall their attacks.
It shouldn’t be too hard, and considering private key authentication, you could even use a single sign-in for multiple platforms/accounts, and use the public key as an identifier to link them across platforms. I know there’s already a couple proof-of-concept Gemini forums/BBSs out there already. Maybe they just need a popularity boost?