Now this is interesting, I know about Tor ofc, with all problems surrounding it (exit nodes etc) but I guess an onion website could be made well protected and shared & updated. You have to host it yourself though I guess.

Freenet, gotta dig down and see how it works under the surface, it looks very promising but it’s kind of complex and I haven’t yet figured out if it is all benevolent sharing for example and what happend if some random node sharing your stuff goes offline.

Very interesting!

I think (I’ll dig more to see if it stands) my advantage would be the redundancy (so the data always stays up and is hard to take down), the no need of benevolent nodes, and potentially the ease if use.

Thanks!

Interesting!

I poked around in the (slightly verbose) documentation and stumbled onto this:

Servers should not re-use URIs, regardless of the mechanism by which resources are created. Certain specific cases exist where URIs may be reinstated when it identifies the same resource,

So I wonder if it has the same inbuilt limitation that IPFS has, which means you cannot just update the data you are sharing, without also having to create a whole new link (I know IPFS are trying to work around that, but have seen no decentralised solution yet).

I’ll poke around some more!

Thanks for the link, I hadn’t heard of them before.

Cheers

Ha ha yeah if there were only more competition :-D

The idea is people run nodes (the hard part: routing incoming internet traffic to your PC) and nodes chit chat (any new nodes? Hey I just changed IP address!, … etc) so that there is a lots if known nodes. Identities rely on RSA keys (like ssh does).

Then on top of that (thus making it useful) you can propose a deal for another node: you share my data and I’ll share yours.

Which means people can access your data from your PC or from that other node sharing it for you.

Share it with several nodes (you obviously are sharing theirs) and there will be a very high probability your data is accessible all the time.

The sharing is completely trust-less, if a node stops sharing your data, you just stops sharing theirs and gets a new partner, no hard feelings.

Added bonus is that it might be shared all over the world, so hard to take down.

All traffic and data is encrypted, so no node even knows what they are sharing.

You can change IP:port address, and update your data easily (that was the big work to be fair).

That’s about it, the implementation is written in python3 + cryptodome and uses RSA & AES-CTR.

The basic use could be to host a website, or a chat for example.

What do you think?

I’m building a sort of new internet protocol, so that we can do away with registrars and dns servers (so everyone can have their “web site”). But I’m quite abysmal when it comes to get people interested :-p

Hello cafe !

The project is on ice atm, if I get the hang on the API calls for communities and messages I’d probably continue.

I know some people think you shouldn’t scan messages, but they’re free to see out there in the open already? I mean a search engine could actually do wonders for unknown communities, and I’d gladly filter out by some tag or whatever if needed. I mean if I get down to do it :-)

Cheers!

Doesn’t work for me (lemmy.mindoki.com) okay I maybe have only one or two users but still 😅. BTW I had to tilt the phone to get put of the too small “reactive” page so that the search bar showed up.

I think you don’t understand how federation works.

It’s like you show something sensitive on TV, and you want to “erase” that from everyone seeing it.

Lemmy isn’t centralized like Reddit or Facebook.

Impossible to be done if not every servers plays by the rules.

Sort of non news too, “don’t put sensitive data on display, especially on the internet”.

Thanks!

I gave a crosspost or two that doesn’t seem to have worked, will rhe come alive :-) ? Anyways I’ll check if a next one works as expected.

Will do and thanks so much for the support 😊💖!

Upgraded and it works well it seems, thank you!!

Thanks for the feedback!

I had some massive problems with the first 0.19, which eventually was fixed but 0.19.3 just was so slow (and only had “minor fixes” IIRC) so after the painful time of the broken federation I reverted to 0.19.2 just because it “worked”. Maybe I should try the 0.19.3 again.

I do have a beefy PC with a good connection and it runs 24/24 so I’ll definitely try out the latest version (again).

Cheers

Yep but I don’t really think small instances work well just yet.

I have a powerful PC (enough RAM, cores, nvme etc on a 1Gb line, 700Mb up) and it’s incredibly slow most of the time, because I guess, there are not many users forcing the federation.

I’m working on some band-aid scripts to half-fix that until Lemmy catches up, but just so you know it’s not perfect for small communities yet it seems.

I don’t know as the (cURL http) API I’m using is confusing for messages so I only have the servers & communities.

If I get it up and running I’ll tell :-D

I made a crude lemmy scanner, but as I’m using the “http api” I’m a bit stuck at servers and instances. Almost 1800 servers and IIRC (I’m on my phone) some 40.000 communities. Yay!

I know the official api docs, but I don’t know how to really use it, just send a json to some server? Would love to have an easy example and I’ll go from there (family, work, art, and other gets in the way).

The goal would be to figure out how to make a search engine without storing the whole lemmyverse locally.

FOSS obviously.

Nice.

I thought it was someone making a bunch of bot accounts for the future…

Thanks for the link, and yeah I devoured the books back in the day but wouldn’t recommend them or talk well about them today because the writer went apeshit pro TERF and more. I mean WTF.

I second that, happy to hear Gabe is still around, gotta head over there!

Hmm… I’d only index public data (I’m not totally there yet of course), which can be found by anyone, but if there is some way for people, posts, communities, servers, to opt out then ok. Serious question: is there though? Second question, I’m wanting to do this open source, which means anybody can take it, remove the check and scan everything. What are your thoughts about that?

If you know how to query servers, communities, posts or comments on that topic I’m all ears, I’m only doing 50% of that today BTW.

On a side note, where is your 0xCAFE come from? Is it like the stack overflow/ memory error checks like 0xDEAD(or 0xDEADBEEF) and so?

Cheers!

Made a little script and raked almost 1800 servers and over 40.000 communities.

If someone has the http api call for messages and posts I’d love continuing and maybe set up some sort of search engine… Or maybe I should go with the soap(?) one but I don’t know how to do it in python, any information greatly appreciated!

Here, have an apple my neighbour!