What are the limitations of DSpace search?

DSpace search uses an index of the content of submitted metadata for both local and federated catalogs. DSpace configuration allows full-text searching of certain document types, including PDF (if text-based or OCR-ed), MS Word, plain text, and HTML. Indexing for full-text searches is not instantaneous--the index is created periodically, usually on a daily basis, as per configuration. DSpace can also be configured to transform and extract text from certain other document types, such as Powerpoint. Some image types (i.e., JPG) can be transformed and extracted to thumbnail images for efficient searching. Potentially less common DSpace document types (e.g., compressed files, .odt, etc.) are limited to search of the corresponding metadata. 

From the user's perspective, DSpace search is effective where quality metadata has been provided for a document and for text-based documents once the full-text index is updated.