NEWWWZ is a reading interface plus an optional perception layer.
This page explains how items enter the feed, how duplicates are handled, and how reactions become aggregated signal.
SOURCES
We ingest stories from selected publishers and feeds (RSS/API), then normalize title/time/source.
- Source transparency: every row links to the original publisher.
- No rewrite by default: the source remains the source.
DEDUPLICATION
Many outlets publish the same story with small title changes. NEWWWZ reduces repetition using:
- Similarity clustering on headlines + entities (people/places/brands).
- Source merging (optional): one “main row” can represent multiple outlets.
- Goal: fewer repeats, not less truth.
CATEGORIES
Each item is assigned to one category using taxonomy rules (see Taxonomy page).
We optimize for readability: users should find what they expect, fast.
AREAS
Area filters (WORLD / continents / countries) are derived from story context (locations, institutions, and publisher metadata).
When uncertain, we prefer WORLD rather than mis-labeling.
REACTIONS → MOOD
Reactions are designed as
micro-signal, not comments.
Aggregations can be displayed by day, area, and category to form a “mood index”.
- No public debate layer inside the feed.
- Aggregated insight over time (reports + archive).
ADS AS CATEGORY
Sponsored content lives in a dedicated ADS category and keeps the same grid.
It’s readable, clearly labeled, and never disguised as news.
Method is a promise: clarity first, signal second.
If we ever have to choose: we protect reading quality over growth hacks.