Skip to content
Merged
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
34 changes: 33 additions & 1 deletion docs/VISION.md
Original file line number Diff line number Diff line change
@@ -1,6 +1,6 @@
# Zeta β€” Long-Term Vision

> **Status:** round 33 v7 after Aaron's sixth pass of edits.
> **Status:** round 33 v8 after Aaron's seventh pass of edits.
> Aaron is the source of truth; this document changes freely.
> The `product-visionary` role (to be spawned, see
> `docs/BACKLOG.md`) will steward it once it exists.
Expand Down Expand Up @@ -91,6 +91,34 @@ Concretely:
- **LINQ + SQL + F# DSL all on the same surface.** Consumers
pick their front-end per-query; Zeta routes everything
through the shared IR + operator algebra.
- **First-class event system.** Not a library-on-top;
subscriptions/projections/retraction-aware streams
are native. `services.AddZetaEvents(...)` wires them
into DI alongside the database.
- **First-class caching system.** The retraction-native
algebra means cache invalidation is free β€” when the
underlying data retracts, dependent cache entries
retract with it. No TTL guessing, no cache-stampede
mitigation, no read-through/write-through/write-
behind taxonomy. `services.AddZetaCache(...)` gives
you a retraction-aware cache on the same store.
- **Cross-API GraphQL with automatic spill-over.** A
GraphQL layer over Zeta β€” cross-service queries
traverse the graph, and because events + cache are
retraction-native, updates propagate across API
boundaries automatically. No manual cache-
invalidation-on-mutation plumbing; no subscription-
pushing-over-WebSocket glue per-field. The graph is
live by virtue of running on Zeta.
- **Pluggable log back-end.** Kafka, NATS, NATS
JetStream, and Zeta's own native log all speak to
the multi-node substrate. Aaron round 33: "all with
pluggable wire protocol, hello kafka, nats, nats
streaming … and our own of course." Single-node
consumers pick in-memory; multi-node picks whatever
log infrastructure they already have; greenfield
multi-node can pick Zeta-native and get features the
emulated back-ends can't express.

What lights up for .NET consumers:

Expand All @@ -104,6 +132,10 @@ What lights up for .NET consumers:
travel queries for free, not bolted on.
- **Distributed-systems research** β€” a reference
implementation of events-as-truth at scale.
- **GraphQL federated APIs** β€” retraction propagates
across service boundaries automatically.
- **Caching-heavy workloads** β€” cache invalidation
becomes a solved problem, not a folklore one.

## Product 1 β€” Zeta the database

Expand Down
Loading