10 (funny) years of Apache Lucene hacking

Search
06/17/2019 - 14:00 to 14:40
Maschinenhaus
long talk (40 min)
Intermediate

Session abstract: 

Berlin Buzzwords 2019 celebrates its 10th anniversay and every year Uwe presented a talk about new in Apache Lucene/Solr. Some of those contained funny details about heavy committing, bugs in certain Java versions, enormous testing and understanding garbage collectors. This talk will present the details of 10 years Apache Lucene/Solr development starting with Lucene 2.9 / 3.0, released around the first Berlin Buzzwords conference. It will walk through the cool steps that led to what we have today like the start of numeric queries, merge of the Lucene and Solr projects, near realtime search, tokenstream attributes, flexible indexing (codecs), finite state automatons, doc values, bugs in Java 7, continuous hacking around memory mapping, force merge confusion aka optimize, randomized testing and code verification with policeman's tools and introduction of Solr cloud.

This talk will not go too deep into details - but as always - Uwe will try to speak as fast as possible to give you an entertaining heavy committing policeman session!