Killing Reticulum to Save It

2026-07-31 · Lew Palm

Mark Qvist put 10 years of hard work into the creation of Reticulum, the only fully-featured stack that empowers us simple mortals to build big-tech-free computer networks, grassroots style. The code is well-thought-out and carefully written. I want to replace it all with Leviculum; that's not a port, but a completely compatible network stack, library, ecosystem, server and toolset.

But what the fuck, Reticulum-the-Python-library works well; why reinvent it? Mark explained that the ecosystem needs cool apps the most, not another foundational library. I agree that the apps are the goal, but we need a new foundation first.

Technical Reasons

We need a lib and tools that are natively compiled, not running in an interpreter / emulator / virtual machine. Programmer convenience (Python's interpreter) comes with the cost of additional layers of complexity. Our mesh tech should get rid of all unneeded layers, be as simple as possible.
Our core lib should be able to run on embedded hardware: the same code that drives a server daemon should run the transport node on cheap operating-system-free hardware. That is also only possible with a compiled 'systems' language; the strongest argument against Python.
Creating mobile apps (for Android) on top of Python is a pain. Python drains the battery on Android.
Mark's UI tools (NomadNet, Sideband) have an unintuitive UI for most users. Mark obviously built them for himself, not for widespread use, which is totally fine. But I want to support programmers to create user-friendly apps, and to create them myself. Apps where you don't need to memorize the keybindings or to learn how the underlying mechanics work before you can use them. That's why I built the lnomad browser and the lblogd blog server, and I want Leviculum to become a smooth-to-use foundation for all high-level app development.

The Structural Reason

The solo programmer of Reticulum, Mark, is our single point of failure. That's always a risk, and it means technical decisions rest on one set of shoulders instead of being community-led. The current model has an inner tension: the technology is built for total equality between the computers, but the development process is centralized, and the reasoning behind decisions isn't always visible from the outside. I feel that Reticulum as a community should share that load, take the weight off his shoulders, and find better ways to decide in which technical direction it should develop. These ways should include as much discussion and transparency as possible. I try to find an open cooperative development process for Leviculum.