Abstract: A stored document is converted into a “web-friendly” format by dividing the conversion job into chunks. Each chunk specifies a subset of the document pages to be converted. The result of processing a chunk is a set of “web-friendly” page images. The page images are stored in a repository and are available for viewing at the client. The remaining chunks are processed in a similar manner. The perceived performance of the system is enhanced because a chunk can be processed much faster than the entire document. If the user happens to advance to a page in the document that is not yet processed, a real-time request is made to process a particular subset of pages that includes the page advanced to.