Have you read http://wiki.apache.org/jackrabbit/DavidsModel ?
Cheers, Alex On 09.06.2012, at 02:24, Günther Schmidt wrote: > Hi, > > this question isn't specifically Jackrabbit related but more on how to > use your JCR, ie. how to organize your data within it. > > I have most of my data now in Jackrabbit now. With data I mean mostly > emails and documents (PDFs, docs, images, etc, calendar and work flow > will follow later). > > The next step for me is to organize it. Ordering the data merely on a > hierarchical basis, ie. "folders" seems wasteful, I could have used a > file system for that and all I would have gained by using a JCR would be > versioning and automatic full text indexing. > > I want to be able to group related content (documents and emails at the > moment) semantically together. At first I had considered using > references to "link" related data but a) it seems difficult to implement > and b) most of what I read suggests not to use them, or use them sparsely. > > So now I'm thinking about using tags instead, ie. add a multivalued > string field and patch it onto an email node or a file node and tag it. > This way an email could belong to more than category. Would it be better > to create a "taggable" mixin for this instead of a plain property? > > Now the tagging approach of course has some drawbacks too depending on > how fine or coarse grained I will eventually use it. If I use it too > excessively by slapping too many tags onto a node item or by using too > many different tags it will become confusing very soon. > > At the the end of the day when I need to take care of business I want an > easy way to have all the relevant data about a case in one place, is the > tagging approach a feasible way to go? > > Günther
