On 8/16/07, Kurt Mackey <[EMAIL PROTECTED]> wrote: > > No, but something along those lines would be really nice. I could get a > lot of use out of Solr, but have no desire to make an environment more > complex by adding in Java.
This is a totally valid assessment. However, there is a lot of work that has gone into lucene and solr on the java platform, and porting it would be a colossal endeavor. Solr is about 50% lucene, 50% operations. The operational aspect of Solr is *very* solid; it's not zero administration, but it's much better than the alternative (rolling your own). We put together a C# client called Solrsharp that interacts through the http accessors to solr. We implemented solr with a java container (tomcat), and it runs as a self-contained system. All of our interaction with it is through the library, which gives us the advantage of working in our chosen framework while leveraging the efforts already being put forth into solr. To Kurt's point about adding to environment complexity by adding Java, Solr minimizes the java exposure. In our operation, we've not ever had to adapt the Solr source code, only modify the configuration files (XML-based). If you stick with what the community has released, you're in great shape. Solrsharp - http://wiki.apache.org/solr/SolrSharp -- jeff r.
