Hey!

With the recent popularity that this subreddit has accomplished, more technical people have started looking at our previous ideas with critical eyes. Several explanations have been published pointing out the problems we'd have to overcome to make our "mesh internet" a reality.

We could just ignore all those facts and continue trying, but there are problems with the following things:

• Bandwidth: by definition, and using cheap radios (half-duplex), the bandwidth would be very scarce. Most users would try to use the meshnet to surf facebook and turn it down after experiencing browsing like it was back in the '90s.
• RF unreliability and randomness: people far smarter than I have already worked in deploying wireless mesh networks, and they have found that the 2.4GHz spectrum is... funny.
• IP Address asignation: though there have been some breakthroughs that could make auto-assignation somewhat possible, there are always risk of collision without having a central authority.

And many more. However, this doesn't mean that creating an internet like the current one with user-owned infrastructure is impossible. It just means that we'd need to improve our technology by orders of magnitude, and it'd be far more costly than exploiting what we already have to our advantage.

If we take a step back, we'd all agree in the following: we want to create a tool to route around censorship. If this is our goal, maybe we don't need to create a new internet. Maybe what we need to do is create tools, hardware and applications which ensure free speech.

My idea is the following:

Remember UUCP? It was popular back in the day. With it, you could route packets to other compuers. Usenet was basically a meshnet, but with dial-up.

Maybe, just maybe, we can create a new layer 2 protocol which doesn't require IP address, and is broadcasted to the whole network. This way, bandwidth wouldn't be a problem: only text messages would go through the network. Maybe small images too.

This would be the perfect way to create some sort of "twitter" for events like OWS. There is no central server hosting it.

This is how it would work: users would have an application in their smartphone which implements our protocol. Let's say we're building a distributed twitter.

When I send a message, I'm broadcasting it over WiFi to nearby people or computers with our software running. It has some metadata, like a uniqueid for that message, the original application used to send the message, and so forth.

Then, my friend, whose phone is close to mine, receives that packet and his application shows my "tweet". Then, he can choose to relay that message, so more people can hear me.

It'd be also pretty nice if we wrote some software for our hardware which would store the packets it receives in some sort of storage, and keep broadcasting them locally. This "repeaters" could be connected to other repeaters, and so on.

I would love to hear criticism about this idea, but I believe it is what we need to do to move forward.