Sphinx is not the best contender for integration, because of it's limited support for incremental updates. It is, however, a good boundary condition on how to design the Indexer API so that a wide range of search engines can work with CouchDB.
This is all still academic, but Lucene seems like the best fit for lightweight integration, with Solr and Sphinx providing a wide range of target support. Eg. if the API can support them, it'd be hard to imagine what it couldn't support. If the glue code for Lucene / Solr / Sphinx can all live on the other side of CouchDB's stdout pipe, then CouchDB will likely supply just enough information through that pipe so that determined integrators can set up any kind of glue that consumes those events. As we explored earlier, the glue needn't be only for full-text indexing. A real-time notification system could use it as well. Maybe the full-text design documents could be used to filter which documents and content get sent over the pipe. Or it could be as simple as a stdout pipe per view function, triggered with each view emit. Then integrators would have lots of flexibility about what to trigger with the events. Chris -- Chris Anderson http://jchris.mfdz.com
