I've been asked a question from "Anirban Datta" from "Tata Consultancy Services" about separating indexing of content based on document types, which means during indexing we want to not index all documents, just some documents based on document types.
By default you cannot do this, but I think using customization we can do that, in UCM we have a service called "START_SEARCH_INDEX" that service is executing a java function called "buildSearchIndex", so If we redeveloped that function again for that service (using reverse engineering), we can make indexing based on document types.
I wish I was helpful for "Anirban Datta"