rns.recipes

Hello, I hope you’re doing well—don’t forget to drink some water! :) I’d like to thank everyone for their efforts—RNS, LXMF, Columba, MeshChatX, and everyone else who has contributed to building these tools. // Well, I tried to communicate with someone on this network using LXMF. I used Columba; their messages would come through, but mine wouldn't go out! I asked them to perform an "announce," which they did, but it didn't make any difference. // We also tried using Sideband, but that didn't work either—though that might have been due to its complexity. // Then we tried MeshChatX. I asked them to add a server, and they were connected, but it still didn't work. One message might have come through by pure chance, but nothing else followed. I asked them to perform an "announce"—which they did—and I did the same, but it didn't change anything. // For the record, the connection was active; for instance, they were able to access the NomadNet site. I am also fully connected—I can ping various "announces," usually with a latency of under 500ms. // So, why can't this person and I establish a connection? Also, since we are in Iran, could this be due to internet disruptions? Reticulum itself appears to be connected, but I wanted to mention this because, despite all my attempts, I haven't been able to communicate with this person on the network—not even with a delay; it simply hasn't worked at all.

This sounds exactly like my first tries to get things work. You might want to read the answers I got - here, for example:

9ce92808be498e9e05590ff27cbfdfe4:/page/forum/thread.muʼcat=help|thread=noob-questions-setup-and-usability

In my experience: don't expect that everything works out of the box. Stick to one program, learn about the technical background (consult the manual), and look into the program logs (most support to view logs from inside the program). For lmxf messages, it makes a difference with which method you try to send (can be toggled from within the program). If you are able to establish a direct link conncection, the other side should receive your message more or less instantly. If you don't have a direct link established (the program tries this automatically for you) and need to use propagation: picking a propagation node randomly might mean you pick a node that is not functional. You might want to find a reliable propagation node and configure your program to use it. And the interfaces you use to connect of course have an impact on the quality of your connection. In theory it is possible that your problem is related to your location, but, as I said, so far what you said doesn't sound special.

Thanks for the reply; I went and read the information in the link you provided. I found it strange that, even though the other person and I were both online at the same time, no messages were getting through directly. As for the person I mentioned—with whom I couldn't manage to have a conversation—they have now left entirely. However, I’ve been talking to someone else over the past two days; the connection was perfectly stable whenever we were both online, and there were no issues with messaging. It’s puzzling to me why I couldn't establish a connection with the first person, yet I was able to chat easily with this other individual—even though both were connecting via the internet.

Yes, I think this is possible and to some degree expected. From what I have learned since my start, one has to keep in mind how the reticulum net organizes itself: If you send an announce (which normally happens automatically when you start your program), it travels from node to node, and this does not happen instantly. I think waiting a bit after you have started up may help. Maybe in your case the second person had a much larger hop distance to you than the first one. As I said, all of this depends on your interface settings (and on the settings of your contacts).

All of this is (necessarily) not trivial. In nomadnet (the program) one can now display the known paths. There are low-level tools to check paths (rnpath; rnprobe). What sometimes can happen is that a node gets offline or has a too large load, and your instance still try to use the respective path. In such cases paths can be manually discarded, etc. - I'm myself still learning how all of this works. I think starting from the moment when I gave a bit love to my interface configuration, and kept rnsh running most of the time, things improved.

And yes, in Iran, nodes getting offline might indeed be a larger problem (there is no "I have no power, going offline now" message...). Anyway, please be careful, don't forget to drink some water, too. And don't hesitate to tell how your experiments are going.

It could be exactly what you mentioned. For instance, MeshChatX doesn't stay active in the background if you close it; perhaps the user kept closing and reopening the app. Or, regarding the connection—switching between a VPN and a regular internet connection—maybe they kept toggling the VPN on and off, or closing the app, which disrupted the application's operation. On the other hand, Columba stays active in the background quite well; we tested it extensively, but it simply didn't work in that case. However, with the other person I mentioned, the connection was established easily—we started on MeshChatX, then they installed Columba, and we were immediately able to communicate over the network (for example, they told me they were switching to Columba, and we connected right away). The first person I mentioned—who was coming from Telegram and had no prior knowledge of this type of network—received a lot of guidance from me; we spent hours trying to get it to work. Could it be that their habit of constantly closing/reopening the app and toggling the VPN caused the disruption? Also, another potential issue with using a VPN is that it might connect the user to nodes that are inaccessible from within Iran. Then, once the VPN is disconnected, the client might connect to different nodes, potentially causing routing issues. I’m not sure exactly what they did; these are just my guesses, assuming there isn't some other underlying problem. Thanks a lot. To be honest, I wanted to bring this up because, you know, if you suggest someone use this network and it doesn't work right from the start, they’ll quickly conclude that the network is completely ineffective. I just wanted to figure out if the issue was on the user's end—or if it was a software problem that could be fixed. :) Regarding routing and such—sure, it can be done manually; for instance, MeshChatX has an option for re-routing. But in my view, this should really be handled automatically. We might be discussing routing technicalities, but an average Telegram user isn't going to think, "Oh, the path is broken, so I need to manually select a different route." That said, I have heard some things about routing; I believe automatic path discovery—in the event of a disconnection—has actually been implemented. Unfortunately, the person I couldn't connect with is no longer around, otherwise we could have done more testing. There’s only one other person I have a solid connection with. :) So, I think that’s where things stand for me—I’ll consider this matter closed for now. :) Thanks again, and take care! And indeed, if my writing has a shameful flaw, I wrote it with a translator.

Regarding routing and such—sure, it can be done manually; for instance, MeshChatX has an option for re-routing. But in my view, this should really be handled automatically.

I don't know if this happens often at all. So far I only saw this problem (breaking path) when browsing nomadnet, never while exchaning lxmf messages. And as far as I know, such issues are being worked on. Things still develop and improve.

Two more comments (your texts read fine by the way):

When I had installed sideband for the first time around three months ago, the defaults "created" a local network around our wland router only. There was no connectivity into "the internet" at all until I added interfaces manually.

Second, this sentence from the manual (section "Understanding Reticulum"):

In general, even extremely complex networks, that utilize the maximum 128 hops will converge to full end-to-end connectivity in about one minute, given there is enough bandwidth available to process the required amount of announces.

It’s good that it’s readable; I kept thinking, "Oh no, this turned out really badly!" So, it’s actually very fast—provided there’s enough bandwidth. We definitely waited much longer than a minute.! (: It was strange; I would tell him to "announce," and he would tell me to "announce," but no "announcement" ever took place. Even after a few minutes, nothing happened. I wish that person were still available so we could test it again and figure out the problem—but unfortunately, he isn't. Thanks again.