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
17 changes: 17 additions & 0 deletions docs/planning/memory-architecture-discovery.md
Original file line number Diff line number Diff line change
Expand Up @@ -374,6 +374,23 @@ The discussion should distinguish:
files, relevant memory, decisions, and provenance without requiring the
original provider transcript.

## Historical Reference: LitRev 2026

The historical [LitRev 2026](https://github.com/yaacovcorcos/LitRev_2026)
project contains a substantial implemented memory and retrieval system. Scient
will use it as a reference implementation from which to retrieve and
selectively adapt useful ideas, not as an architecture, dependency,
compatibility target, or system to copy wholesale.

Useful reference areas include capture-before-belief, scoped memory,
provenance and authority metadata, review-before-promotion, lifecycle controls,
bounded and auditable retrieval, and durable background extraction. Its limits
are also evidence: Scient should avoid treating ordinary event history as
memory, injecting every user preference globally, duplicating canonical
project records, or inheriting storage-specific assumptions. Every borrowed
pattern must be reevaluated against Scient's local-first, portable,
user-controlled, and future-cloud requirements.

## Persistence Questions Come Later

Only after the scope, authority, lifecycle, privacy, portability, and sync
Expand Down