This is a very "noobish" question, but I'm working on a project with a document database and I think they are using it wrongly. Thus, I kinda want to confirm this, as I feel it is tainting my experience.
In a nutshell, they seem to have brought a relational mindset to a document paradigm. For example, instead of embedding a particular JSON object in another, they are pulling them together in the business layer via ids (smells very much like foreign keys to me).
Now, it's true that repetition of data is avoided here (hey, as a SQL guy, I'm all about 3rd normal form), but I thought that having the data in there as part of the bigger document was more important than avoiding repetition of data.
I get that there are probably trade-offs and it may not be wise to have huge objects that contain everything, but I was hoping for a bit of guidance as to what is the best way to do things in this space and whether this place (probably) has it it wrong. (NoSQL is challenging if the blind is leading the blind).