@codemonkeymike I've even found people interested in my obscure asynchronous communication interests like #NNCP and #UUCP. Twitter was never great for less popular things and got worse over time as it tried to boost the popular.

#Mastodon is the closest thing I've seen to a 90s #BBS, and this is a compliment.

You can also follow hashtags here, so any post by any person that is followed by anybody on your instance with that hashtag will show up in your feed. 2/

@kyonshi This is one reason that I mostly use #NNCP instead of #UUCP these days. There are some use cases where UUCP is *perfect*. For instance, long-distance, low-power, low-speed radios such as #LoRa or #XBeeSX. LoRa has speeds similar to the modems UUCP used back in the day, and it can be quite useful for that. But nowadays, nicer support for USB drives and encryption is a win.

@kyonshi Welcome! I was sort of half-consciously scrolling along, and then all of a sudden my brain went, "Wait, what? Did I see #UUCP in that list? Why yes I did!" Hello, fello human that knows what UUCP and Usenet are! I have been running UUCP (and its successor #NNCP) off and on for 30 years. I have a page at https://www.complete.org/uucp/ and there's https://www.complete.org/nncp/ about NNCP. Curious what you are doing with UUCP and who you connect to. Also I offer Usenet feeds via NNCP.

@kyonshi Yeah, it's from an earlier era. It is amazing over low-bandwidth and lossy links (LoRA radios, for instance) but isn't entirely at home in 2024. #NNCP its its spiritual successor (sort of like ssh is to telnet), and I use it a lot. More modern setup, everything encrypted, etc. https://www.complete.org/nncp/ has some links. NNCP doesn't deal well with noisy serial links though

I'm curious what you're doing with your UUCP setup.

@kyonshi That is, in the port config file, are you using tcp or pipe? And if tcp, is port 540 open? (tcpdump -n -i interfacename port 540 on each system to look at what's going on)

Hope that helps! Let me know! Also if you are interested in #NNCP I can help with that too. https://www.complete.org/nncp/

@oh2cil @KC1PYT @TwiceShy I make that distinction because Usenet can flow over at least three protocols (NNTP, #UUCP, and #NNCP) and the rnews interface is simple enough that it could easily flow over more.

NNCP has a "multicast areas" feature which is sort of a generic replication of the Usenet distribution model.

I should also note that #NNCP is always 100% encrypted. This makes it awesome for a lot of use cases, but problematic for ham uses.

@oh2cil @KC1PYT @TwiceShy Hi! You have piqued my interest. I enjoy #NNCP, run a #Usenet transit server, am a ham, and have also written code for #LORA
(I'm not the #NNCP author, but I talk about it a lot more than the author does )

If you're talking about something to exploit the fact that you can reach multiple radios on the "RF broadcast domain", yes #Usenet would fit the bill well, but #NNTP would not, since it's a 1-to-1 TCP connection.

@TwiceShy We should compare notes! This has my eye: https://nncp.mirrors.quux.org/. I know I've talked to the author of #NNCP but I'm drawing a blank on his name.

@bplein So I want something that can handle somewhere around 2-4 SATA drives and has 8GB or more RAM. It is a backup machine that receives backups using #ZFS or #dar over #NNCP

Currently I use USB "toasters" but I am having reliability problems with them, and figured that something that is native SATA rather than USB may be better.

I could see possibilities for future use involving a small VM and such, so being able to go above 8GB RAM is a nice to have. Does that answer it?

@ariadne @viq At the same time, this looks like yet another server-centrism issue to me.

You should, as an instance, be able to refuse to federate with anyone. Yes.

However, I should be able to connect my client to however many instances at once that have a copy of the room and rebuild the history client-side if I so feel like it. I should also be able to unilaterally ignore/block everything from instances or users I don't care to read from or about.

I should be able to add new posts to the room while offline and resynchronize them up to various instances at some later point when I get online. Or I could send them to some instances ahead of time as #NNCP messages.

@jgoerzen @kelbot @ajroach42 : I’ve read a lot of your stuff about #NNCP when trying to see how we could decentralize #gemini
(see https://ploum.net/2021-10-10.html and https://ploum.net/2021-10-25-offmini2.html ).

So yeah, there’s clearly something that could be done here.

With the soon to be released Offpunk 2.0, the code has been split in several independents CLI tools. All the networking/caching is done by "netcache" and I would happily discuss what could be done to integrate it with NNCP.

@ploum @kelbot @ajroach42 Oh great! Thank you! I do have a question. I have been thinking "this would work great with #NNCP." As in, if I ran offpunk --fetch-later or --sync, it could generate a little command file that could be sent asynchronously to a machine with Internet, and then it could respond with the requested data. (apt-offline does something similar for Debian repos). I thought maybe I could make this work by syncing its cache dir or something. Is that realistic?

@irenes @BoydStephenSmithJr #NNCP is somewhat similar so maybe it can be of interest.

@screwtape @ellenor2000 Not necessarily per se, it's really only once one has both #P2P and #AsynchronousCommunication that things get very interesting.

Unreliable uptime also isn't quite the asynchronicity I refer to (https://www.complete.org/asynchronous-communication/). Rather, I'm referring to a system developed with no expectation of ever having a low-latency route to the end destination at all.

#Usenet is your classical example. Use #UUCP (#NNCP now) or mailed floppies to update your spool. It's supported.