Althea started out with a similar pitch. I'm not sure where they are now

https://althea.net/how-it-works

just an lurkers prospective, but they really seem to be growing fast. They have a lot of networks up and running and have been doing a pretty impressive job of getting people real broadband internet where there wasn't any before.

We're hiring.

https://twitter.com/AltheaNetwork/status/1357794950843305984

https://althea.net/careers

I have an idea for a bunch of low powered nodes that make an sms analog, a bbs and an ebook library available during natural disasters. It'd be nifty to allow communities to communicate during extended disasters, where 4g towers can often lose their backup power.

I would love it if one of these mesh wifi projects really made it somewhere. That would be very convenient for this project.

Remember the guy who rescued a lost hiker by deducing his location Sherlock-Holmes-style using a single photo? [1][2]. The story was trending on HN few days ago.

He's got a pinned tweet [3] that says "One of the key technologies which has helped in crowdsourced response to extreme weather disasters (fires floods hurricanes) these last years has been ability to monitor local emergency radio channels remotely. Let's get these everywhere. #raspberrypi #rtlsdr #scanner"

He (Benjamin Kuo) is a HAM Radio operator active during disasters.

It's different from what you wanted, but just putting it here.

[1] https://www.nytimes.com/2021/04/15/us/missing-hiker-found.ht...

[2] https://twitter.com/ai6yrham/status/1382371967618097157

[3] https://twitter.com/ai6yrham/status/1085609599661617152

If the frequencies used are above the plasma frequency of the ionosphere (typically >30 MHz) like here then the only solution to "mesh" or cell networks is line of sight and that only comes from height above terrain. And height above terrain, except in rare and unique cases, is always expensive.

The only reason cell networks work is that the telcos pay big money for getting access, power, and backhaul to places up high. No amount of power increases, modulation gain, reflector aperture, software, or anything else you can think up is going to get around the lack of line of sight.

Unless a decentralized mesh network has enough money to do this it will only ever work in specific regions where single individuals or groups can pay $$$ and cover large areas. Like in Seattle where the mountainous coastal terrain is useful for hamwan's network, or in mega-cities like New York where residents have skyscraper roof access for p2p wireless community ISPs.

Actually, wireless ISPs are very common around the world. Yes, line of sight to another node or a tower is always necessary.

There's such an odd divide between tech geeks who like to sit around and think about "p2p wireless decentralized mesh networks", and don't know of any network that doesn't use those buzzwords, and wireless ISP operators who actually build them in the real world despite never having heard the word "mesh".

Generally the networks started by the mesh geeks perform much worse, and have few, if any subscribers, despite the fact that topology and equipment are identical to a wireless ISP.

Source: was a mesh geek, started a company making open source software for p2p mesh networks, now it serves wireless ISPs, who tend to have 100x the hustle and professionalism of the mesh geeks.

That's very interesting. Do they deliver internet through WiMAX ?

A wireless ISP is not running a "mesh". But yeah, having an income stream to pay for height above terrain makes all the difference.

Some run meshes, some don't. Whether or not they use a distance-vector routing protocol is a pretty insignificant technical detail IRL.

They're good solutions for places like Athens where the population is dense and urbanized and people are more concerned with getting any connectivity at all rather than getting low ping times in CSGO.

Is this like an open-source Find My?

I feel like every such project should be asked to justify its existence by explaining why it's not actually identical to some other similar project. Reviewing the existing literature might save a lot of repeated work.

Even if the answer is "the community that runs this one is friendlier", hey, that's a valid differentiation, okay. But seriously, there are ten bazillion wireless meshes now, what makes this one worthy of anyone's attention or effort?

Justify to whom? If you put in the research and work to build a thing, you get to have it exist.

Even if there are already similar projects, each next one may end up with a novel variation that could contribute to the broader understanding. Or even if not, even if it's just like all the others, it still has value for those learning from being involved.

To the people these pages are for, doesn't seem like they were built and spread for the author's own reading. Not that "I wanted to build my own to learn" isn't justification as well or wouldn't let people know what's special about it one just can't find such information in the pages.

Helium is taking off in the IOT space. I am not sure if it can be used for regular communication like messaging or not, but that would be amazing if it could. https://www.helium.com/

Mesh is what we can look to when we don't have (reliable) Internet, for when we're in places that don't have Internet, and/or for people who don't have the resources to get on the Internet.

Rust is not something we're going to be running on low power Pi-type boards in trees, or on the kind of hardware that's available to people with very limited resources.

So why write software in a language that isn't meant for and can't be compiled on modest hardware?

Why wouldn’t you be able to run rust on something like a Pi, or even a much more modest mcu? (You totally can)

There’s even a whole section of the rust documentation about embedded development in rust. https://docs.rust-embedded.org/book/

Rust is something you don't want to compile on a low-power device. But it's totally something that will happily cross-compile for a low-power device. Many people use Rust for such reasons as high-performance code and low resource footprint of it.

Frankly, you won't run GCC on an MCU either.

But if all you have is an RPi, you can still run the Rust toolchain right on it, if a bit slowly; e.g. https://opensource.com/article/19/3/physical-computing-rust-...

This sounds like you think Rust is interpreted. It's not.