Can someone explain CouchDB's advantages to me? From the cocktail napkin diagram on their apache page, it looks like the only major advantages over vanilla rep

Can someone explain CouchDB's advantages to me? From the cocktail napkin diagram on their apache page, it looks like the only major advantages over vanilla replicated Solr would be 1) Better replication architecture (much better?) 2) Convienence layers in javascript for querying What am I missing? Is the main purpose of the project the replication or the javascript libraries? Is the fact that you can update to arbitrary hosts as well as read from arbitrary hosts the big win? It doesn't look like it's sharded so I can't see any huge scalability gains over Solr..