Thread explorer classified
Every sampled, classified thread from dev@ and JIRA comments, linked to its public archive record — a Pony Mail thread permalink for dev@, an issue link for JIRA (D27 amends COMMUNITY-HEALTH.md §7.3's "no deep links" rule for threads only; every other §7 rule stays binding: no per-message scores, no per-person data or names, no quotations, plain neutral column names only). Default sort is date, most recent first — there is no severity-based default sort or highlighting.
Method note. This table lists every sampled thread that has at least one classified message. A sampled thread with nothing to classify — a JIRA issue with no comments, or a dev@ thread whose only message(s) came from an automated sender (e.g. a CI bot) — is filtered out before classification and never appears here; this is the large majority of the gap between the sample size reported in the Conversation patterns section above and this table's row count. No thread with at least one classified message is excluded from this table for any other reason.
Threads with ≥1 flagged message, by label, year over year
Share of that year's (filtered) threads with at least one message classified ≥0.5 for the label — one small mini-chart per label, year on the horizontal axis. Constructive and negative labels are two independent-scale groups of panels, and each panel's own y-axis scales independently within its group — same visual language as the Conversation patterns section's own year-over-year chart.
Constructive / discussion
Negative
Counts are of sampled threads (K=60 per venue-quarter); for population-weighted rates see Conversation patterns.
Export the current view:
Loading threads…
| Date | Venue | Thread | Messages | Participants | Outcome | Peak intensity | Flagged messages | Feedback |
|---|