Hi Peter, You need to use Lucene,
- To have more control - You cannot depend on any Web server - To use termvector, termdocs etc - You could easily extend to have your own Analyzer You need to use Solr, - To index and search docs easily by writting few code - Solr is a standalone App and it takes care most of the stuff like optimizing,warmup the reader etc.. - Solr could be extended to multiple nodes - To use facet If you are developing your client in Java and want to use Solr then i would advise to use SolrJ as it is easy and you don't need to care about HTTP stuff. I use Solr using SolrJ in my project www.findbestopensource.com Regards Aditya www.findbestopensource.com On Wed, May 19, 2010 at 4:08 PM, Peter Karich <peat...@yahoo.de> wrote: > Hi all, > > while asking a question on stackoverflow [1] some other questions appear: > Is SolrJ a recommended way to access Solr or should I prefer the HTTP > interface? > > How can I (j)unit-test Solr? (e.g. create+delete index via Java call) > > Is Lucene faster than Solr? ... do you have experiences, preferable with > the same index? > > The background is an application which uses Lucene at the moment but I > hardly need the facetting feature of Solr and I don't want to implement > it in Lucene for myself. > > Regards, > Peter. > > [1] > > http://stackoverflow.com/questions/2856427/situations-to-prefer-apache-lucene-over-solr > >