Current main problem with WQS SPARQL results is that servers in SPARQL cluster may or may not in sync. This causes that user can lose points when query hits to server which is 1+ hours behind from current state.
Examples
Proposed solution is to execute SPARQL queries on both https://query.wikidata.org and https://query-main.wikidata.org/sparql and merge those results. Con is that it could delay removing points, but it would remove wobble.
Current main problem with WQS SPARQL results is that servers in SPARQL cluster may or may not in sync. This causes that user can lose points when query hits to server which is 1+ hours behind from current state.
Examples
https://fi.wikipedia.org/w/index.php?title=Wikiprojekti%3APunaisten_linkkien_naiset%2F2026&diff=23948684&oldid=23948671
https://fi.wikipedia.org/w/index.php?title=Wikiprojekti%3APunaisten_linkkien_naiset%2F2026&diff=23949433&oldid=23949369
https://fi.wikipedia.org/w/index.php?title=Wikiprojekti%3APunaisten_linkkien_naiset%2F2026&diff=23948447&oldid=23948353
Proposed solution is to execute SPARQL queries on both https://query.wikidata.org and https://query-main.wikidata.org/sparql and merge those results. Con is that it could delay removing points, but it would remove wobble.