Hi,
regarding this post: https://hub.alfresco.com/t5/alfresco-content-services-forum/speed-up-community-alfresco-with-lucene-...
Our goal is to make Lucene indexes lean and to remove stop words. We tried to implement custom dataTypeAnalyzers, but the result indexes show it didn't work.
We assumed that nodes will be processed with locale which is in property sys:locale. Then we made a custom dictionaryModel.xml in alfresco/extension model which pointed to custom alfresco/extension/model dataTypeAnalyzers*.properties files which pointed to our custom LocaleAnalyzer class which had among other things stop words to remove.
As I said before, stop words are still indexed. 
