The current implementation of Gallery2's search functionality suffers from poor performance, a problem particularly pronounced for large datasets because indexes cannot be used to improve efficiency. The solution to this performance and scalability problem is to use DBMS agnostic full text indexing for searches. Furthermore, the current search implementation does not support complex queries (such as queries containing boolean operators) or effectively support even basic relevance-based ranking algorithms. The use of a full text index and an enhanced query parser could offer improvements in both these areas.