Hi All, In LUCENE-8841[1], the question of having a relevance benchmark which allows measuring impact of changes to the relevance of common usecase queries and guarding against unintentional regressions.
Adrien very kindly pointed out that Open Relevance project [2] is a good candidate for that goal, and it would be a nice thing to revive it and make it an updated and active suite (maybe with nightly benchmarks on the lines of luceneutil?). This thread is aimed at discussing the potential paths forward: 1) Do not revive the project. 2) Revive Open Relevance Project and make it a sub project of Lucene again 3) Revive Open Relevance Project and propose it as a generic performance benchmark for Apache Software Foundation, not as a Lucene subproject. If we go down the path of 2), I would suggest a working committee that can ensure the continued health of the project, to ensure it does not go down the way it did in 2014. Please share your thoughts. Regards, Atri [1] https://issues.apache.org/jira/browse/LUCENE-8841 [2] https://lucene.apache.org/openrelevance/ -- Regards, Atri Apache Concerted --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
