Roadmap
Every milestone stays runnable.
NexusFS was staged as a vertical-slice project so the repository stayed a usable system at each step, not a pile of disconnected prototypes. All eight milestones are complete. What follows is ordinary product work, and it should be picked by value rather than by sequence.
The order it was built in
- M1 completed persistent local namespace state.
- M2 shipped one real external facade.
- M3 completed peer replication.
- M4 layered in encryption and transparent proofs.
- M5 made replication adapt to power, heat and link cost.
- M6 gave operators collection, versioning and enrolment.
- M7 made single facts about state provable, without overclaiming.
- M8 extended that to absence, so a deletion is demonstrable.
- A real proving system stays behind its own milestone.
What would come next
- Detect metered links per platform, so the scheduler rule that defers content on a metered connection can fire outside a test.
- Wire per-recipient key envelopes into the write path, so replicas need not share one repository key.
- Decide whether a mountable interface is wanted — and if so, pick WebDAV or FUSE on the merits rather than by default.
- Make the Merkle tree incremental, if measurement says the rebuild has become the bottleneck it is not yet.
Companion Planning
The backlog is now split out separately
The roadmap shows milestone outcomes. The backlog page shows the concrete work queue required to reach them.