The classic starting point is David Nüscheler's (JCR spec lead) content-modeling guide:
http://wiki.apache.org/jackrabbit/DavidsModel You should design the data model not based on the application code, but from a content standpoint: what belongs together, what is read/written together most of the time and how can it be organized effectively. The latter can always be verified by how easy it is to browse through the content tree as a human. And the hierarchical data model (together with less-strict typing, eg. nt:unstructured) gives you freedom to change the model later (eg. introduce another entity, ie. folder) - you are not forced to design the perfect model up-front, as you are more-or-less forced with relational database schemas. Regards, Alex On Mon, Jun 2, 2008 at 3:05 PM, Florian Holeczek <[EMAIL PROTECTED]> wrote: > Hi all, > > is there any experience on how a content-centric application should be > designed? What I mean are patterns like DTO and so on. > > I wonder if a user defined mapping of names from internal application > data to content model data is indispensable. > For example, think of the case of a content repository containing wiki > pages and it isn't a priori clear, which wiki application will be used > (or maybe it's clear that there will be used several different ones). > > I haven't been designing any non-trivial JCR application yet, so any > help or hint is appreciated. Maybe this even leads to a discussion! > > Best Regards, > Florian > -- Alexander Klimetschek [EMAIL PROTECTED]
