Hi. I'm a CMP/CMR newbie, trying to make some design decisions, and thought I might ask here to see if anyone had ideas on my problem.
I have a document object with some simple metadata, which in turn owns a bunch (5-50) of page objects. The page objects each have a bit of metadata, and then two 1.4MB binary blob images.
Does it make sense that the page objects be separate beans with CMR relationships, or can the page objects simply be attributes of the CMR document object?
Once concern is simplicity; CMR seems to add quite a bit of complexity. Also, I don't see much value in the CMR approach because the pages aren't really useful except in the context of the document.
But, I'm most concerned with efficiency. I don't want all the page images loaded up unless necessary; some access will only need to see the metadata. It looks as if the attribute approach would result in all the page objects being retrieved together, and each would apparently load the images.