hi toby the usecase is about storing a lot of binary documents (*.doc *.pdf *.tif ...) per "project". filesizes will be from a few KBs up to a few MBs. Every change in such a document done by a user has to be stored as a version. (it is necessary to be able at any time reconstruct all changes in a document and to detect who did those changes) a document will exists at least a few years and the number of versions can grow over 100.
so we are talking about TBs of data and millions of nodes with binary content. therefore performance and used space are two very "hot" topics i have to handle. -- View this message in context: http://www.nabble.com/about-document-history-t1296233.html#a3528700 Sent from the Jackrabbit - Dev forum at Nabble.com.
