Up until now scanning applications I have used have tended to work around the page or sheet as the fundamental concept to each page scanned gets written to its own file and, when a document has multiple pages, the link can be indicated by putting the page files into a directory or naming then in a consistent way. For the page files themselves I have always used PNG which I almost always use for single images that are not photographs.
I've recently been trying a new Gnome application that came with Ubuntu 10.04 called SimpleScan. Unlike previous applications this is document based in so far as you can scan a number of pages (including the whole contents of the document feeder) and then save them together either as a single file or a set of related files (one per page).
This has got me thinking again about files that can hold more than one image so the file then corresponds to a document and the application that views the file presents the various pages as needed.
PDF is very fashionable for scanned documents and this is one of the options available for saving. I note from the Wikipedia article that PDF has now been standardised by ISO but is encumbered by patents which Adobe licenses royalty free. Any thoughts on how good a standard this is for long term storage and whether there are better alternatives?