Files
dotfiles/claude/memory/feedback-bmf-ingestion-rate.md
David F GliddenandClaude Opus 4.8 3f9a89b00c chore(memory): Basic Memory trial begins — sync normalization baseline (283 files)
Basic Memory v0.21.6 first sync over the live memory dir (steward-authorized
live-dir trial, Option A 2026-06-06): adds permalink: to frontmatter, refolds
long YAML description lines, strips final newlines. Bodies untouched —
verified via full diff classification. From this commit forward, any diff in
claude/memory shows only what Basic Memory or the session writes.

Trial design: MemPalace untouched as incumbent; git status check on this dir
at every wrap; end-of-day evaluation (recall quality, sync robustness,
rebuild-from-files, malformed-file behavior).

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-06 09:52:17 +02:00

1.5 KiB

name, description, type, originSessionId, permalink
name description type originSessionId permalink
BMF connector_status rate/ETA is unreliable Don't quote the `rate` or `eta_minutes` from BMF connector_status as a time estimate — ingestion is bursty, rate changes substantially, linear extrapolation misleads. feedback e60f4993-8250-41bb-a95f-c288da448be5 claude-memory/feedback-bmf-ingestion-rate

BMF connector_status rate/ETA — don't trust the extrapolation

Rule: When reading mcp__capablemind__connector_status, do not report the rate or eta_minutes fields to the steward as if they were reliable predictors of completion time.

Why: BMF ingestion exhibits a burst pattern — periods of slow steady throughput interleaved with rate surges. The eta_minutes is computed by linear extrapolation from instantaneous rate, which doesn't capture the burst behavior. Steward has observed this repeatedly across vault ingestion runs. Quoted observation (2026-04-18): "the ETA is not a good indicator because the rate will change with the burst ingestion pattern that seems to take hold."

How to apply: When reporting progress during long-running ingestion:

  • Report processed / total_estimated and entities_extracted (raw counts — informative)
  • Skip rate and eta_minutes unless explicitly asked — or caveat them heavily
  • Frame duration in coarse terms ("several hours" / "let it run overnight") based on total size, not fine-grained ETAs
  • Poll intermittently if asked; don't build plans around a specific finish time