From: nospam@example.net
Sounds interesting! I don't know which way would be the best one to go. To
fork leafnode, and add/remove stuff, or to write from scratch.
If you only focus on a subset of nntp maybe writing from scratch might not
be such a huge task?
On Tue, 25 Feb 2025, Salvador Mirzo wrote:
> Anton Shepelev writes:
>
>> D to Salvador Mirzo:
>>
>>>> The web is like that. A website sends you to another
>>>> one. This is decentralization. No NNTP servers send
>>>> you to another one, except those that have peers, but
>>>> then it's as if they're all the same. My idea is to
>>>> make NNTP servers more like the web.
>>>
>>> That would be a nice take on NNTP. A kind of "federated"
>>> nntp model, as opposed to todays standardize and "global"
>>> version. Federated is perhaps not the right word for it.
>>>
>>> I think I see your vision here... we could think of the
>>> local nntp servers as small communities, you could opt-in
>>> to make them public, keep them private, or just register
>>> them with a search engine if you want.
>>>
>>> That model also would avoid all the newsgroup hierarchy
>>> stuff, you just name your groups what ever you want, and
>>> you can decide to setup peers with other communities you
>>> know.
>>
>> Is the difference from the current Usenet so big? There
>> already are many servers, public and private, and peering
>> with one another. news.tilde.club is an example of a small
>> NNTP community.
>
> It's not too different. It's close. But not quite. For instance, in
> my idea---which I acknowledge that I did not present it properly---, the
> NNTP server is meant for a small community of people who would like to
> have an interesting community. I'm writing ``interesting'' on purpose
> because it is completely vague. But you'll get the meaning of the words
> from the properties of the community.
>
> Which properties? Let's leave that open as well. Let's go straight to
> the implementation of system behavior that has been designed to support
> these properties---whether it will work or not.
>
> In an NNTP server of such desired communities, every member should have
> the same rights as every other member---meaning powers in the system.
> So every member can create whatever groups he wishes. How could NNTP
> systems allow this? Clients don't have a way to send arbitrary commands
> to servers. So the idea of such NNTP server is to allow a remote TCP
> connection to let users interact with the server beyond of what NNTP
> really allows. In other words, the system should be hackable. And it
> might be even literally hackable. For example:
>
> --8<-------------------------------------------------------->8---
> $ telnet nntp.server.somewhere.com 119
> Trying 1.2.3.4...
> Connected to nntp.server.somewhere.com.
> Escape character is '^]'.
> 200 Welcome!
> repl
> REPL> (+ 1 1)
> 2
> LOOP> quit
> 200 Okay, done.
> quit
> 205 Good-bye.
> Connection closed by foreign host.
> $
> --8<-------------------------------------------------------->8---
>
> You get the idea. Honestly, this look rather pointless to me right now.
> (You can use the REPL to use someone's account, say, without the person
> ever finding out.) It's really just a joke in taking ``hackable'' to
> the extreme. I can also say it's a homage paid to the MIT hackers in
> the 50s and 60s, who built debuggers that let you fix problems in the
> system as you used it, stories told in the book ``Hackers'' by Steven
> Levy, 1984.
>
> Anyway, sorry for the tangent. Let's get back. So by way of shells
> like that, you can have commands that let you create groups and create
> user accounts and other things. For instance, we can implement search
> beyond XPAT, say, and many more things for fun or whatever.
>
> So I want to provide people with a tool that's very easy to install and
> lets them create a closed community with a tree of users. What tree?
> To get an account, someone on the inside must create your account and
> then the system records has invited who. It's clear to everyone who
> knows who, who invited who. This might create a certain sense of
> responsibility in the users. There should be a command that lists all
> users and the tree of users, so that's public information:
>
> users
> 200 List of current users:
> A, last seen on Mon Nov 25 14:40:16 2024, invited (B C)
> B (account locked: disappeared for over 3 months), last seen on Sun
Nov 17 23:39:21 2024, invited nobody
> C, last seen on Tue Feb 25 14:42:24 2025, invited (D E F ...)
> [...]
>
> You get the idea.
>
> This idea stems from my hypothesis that a good community is one that
> unites people in cyberspace, but these people have a real connection
> with one another that goes beyond the mere interaction that might be
> taking place between them in cyberspace. Say I invited you. Then
> that's because I already know you somehow. You then invite someone
> else, whom I have no idea about, but I do know that that person has a
> connection with you, so it has an indirect connection with me. Of
> course that people might just randomly invite one another; it's a system
> with mathematical guarantees.
>
> Just let you understand me a bit more: I came to the conclusion that an
> a community in cyberspace should have a connection in the offline world.
> So I invite a friend from mine school, who might invite someone across
> the world, but this remote person has a connection with my friend, who
> has an external-world connection with me.
>
> And there's much more---I think it's nice to have a cohesive group,
> which could use the knowledge of who is reading the groups, closing
> accounts of people who don't have the interest in participating. This
> allows a person who is writing to answer the question---what's my
> audience? Those that have been logging on.
>
> In other words, there's no privacy. The idea is for a closed community.
> USENET access doesn't interfere with the community because the USENET
> can't read the local groups.
>
> Although there's no privacy, many experiments can be done. We could
> have a command, for example, that enables randomization of names in a
> certain group so that nobody ever knows who posted what. But such idea
> is nothing but a game---it's usually very easy to detect who writes what
> in a small group.
>
> The idea really started out as a playground for programmers, but it has
> evolved to some of these ideas. There's much more to it, but I plan to
> dissertate on it only if I release a first prototype.
>
> If you have any ideas, I'd love to hear about it.
>
--- SoupGate-DOS v1.05
* Origin: you cannot sedate... all the things you hate (1:229/2)
|
