There is a great discussion of this issue over at the PzDb website.

The addition of methodology by users that was never conceived of by the program authors coupled with the total lack of standards for file naming has resulted in a Poser runtime where the naming and location of all content is basically RANDOM! It's not that the Poser Library is lame. Especially since the latest rewrite, which is quite an advance. The problem is, in the native library, there is nothing that ties all the pieces together. Collections is an attempt, but it's sure tedious to do. That's why I went with PzDb's database approach.