---
title: "Comment by Ahmet Özel on \"Evernote Meets Cognee - Without Stale Memories\" | Hashnode"
description: "The saved USN cursor also needs to be associated with the notebook and tag scope that produced it. If a user expands the configured scope, previously excluded n"
url: "https://hashnode.com/posts/evernote-meets-cognee-without-stale-memories/6ae417263055b2aa33ff5519/comment/6ae430b187c1a0f4df01dd87"
---

[Discussion](https://hashnode.com/posts/evernote-meets-cognee-without-stale-memories/6ae417263055b2aa33ff5519)

Comment [Evernote Meets Cognee - Without Stale Memories](https://hashnode.com/posts/evernote-meets-cognee-without-stale-memories/6ae417263055b2aa33ff5519)

[Ahmet Özel](https://hashnode.com/@ahmetozel)

[Oct 30](https://hashnode.com/posts/evernote-meets-cognee-without-stale-memories/6ae417263055b2aa33ff5519/comment/6ae430b187c1a0f4df01dd87)

The saved USN cursor also needs to be associated with the notebook and tag scope that produced it. If a user expands the configured scope, previously excluded notes may not receive new updates and therefore never appear in incremental chunks after the existing cursor.

I would treat a scope change as a reconciliation event, with a test that adds a notebook containing unchanged notes. The opposing test narrows the scope and checks removals only after a complete listing succeeds. That exercises configuration changes separately from the update and deletion events your fake client already covers.
