I have a content type for a collage of six photos. For each photo there is a caption, a link, some link text, and the photo itself. The photos are of different sizes and in different places, and there can be only one photo in each spot per collage.
If I just use Field Groups, I have to create a different CCK field per photo per field, for a total of 24 fields.
If I use a single level of Field Collections I can get that down to six separate Field Collections and four CCK fields, but if I need to make a change that applies to all six photos, I have to make the change six times.
I can use two levels of Field Collections, with a main Collection that contains the four fields, and six other Collections named "Top Left", "Top Right", etc., each of which contains only the main Collection. This is maximally abstracted and would probably do, but it makes the node editing interface look like this:
Those redundant fieldsets are bothersome and are a dead giveaway that the underlying data is architected badly.
Using a single Field Collection that allows multiple values seems the most correct. It gets me down to one Field Collection and four fields. But the repeated groups don't have any identifying information other than their deltas, and if the client ever needs to add one they won't be able to keep track of which line corresponds to which spot in the collage.
Is there a way for me to put identifying information on each of the repeated values of a field or Field collection? Otherwise, is there a better way for me to do this?