Problem

The ChatStack is a complex.

• It has core dependencies on Rust libraries that cannot be ported easily.
• These are exported to C then wrapped in Nim to handle interaction with λDelivery, then converted to C again and then wrapped in QT cpp.
• This wrapping and unwrapping accounts for 43% of all source code in the stack, and that does not include the added complexity from the build pipeline.
• In the Last 2-3 months 80% of the teams effort has been dedicated to “non-chat” features involving the complicated builds, compiler incompatibilities and FFI wrapping.

The team cannot remain effective with this level of complexity in the project, especially as the number of problems that need solutions grows.

Current Architecture

The largest contributor to complexity, is FFIs and build steps to compile across 3 languages

https://www.figma.com/board/nB4Pk9pPx5uIR0xH1yCKDP/Jazz-Diagrams-Central?node-id=653-2173&t=FTG8KFC9CuhntKLT-4

Proposed Changes

Background & Motivations

1. The projects only requirement on nim has been λDelivery. As λDelivery must now be accessed via logos core (written in C). The question becomes how should the architecture change to reflect this?

2. With questions arising about the compatibility of LogosCore in Mobile, choosing a lean flexible model is important to be able to respond to discoveries and learning into the future.

3. Simplifying build process will allow ChatTeam to focus on the new work areas that are emerging.

Repositories and Responsibilities

The proposal is that the repositories take on clear roles.

Libchat - Core Implementation of Chat protocols and cryptography. This avoids Implementation details and focuses on isolating the protocol and cryptographic implementations for review. This should be purely synchronous functions.