If you want/need the newer features I would say sure use BerkeleyDB. I'm a bit of a minimalist, so if I don't need the newer stuff, I like to stay with the simpler module. I probably just stick with what I know better I guess.

Disadvantages - BerkeleyDB doesn't come in the perl core, DB_file does. Scripts which rely on the newer API features won't be directly portable to a machine which doesn't have this module installed. If other people are going to use your code I would stick with DB_File. This is a minor quibble, but it depends on sophistication of your users.

Advantages - the OO modules and the new features. Queues are very nice. Also you can setup multiple dbs within a file eliminating the need to create multiple index files for different datasets. Lots more things, really depends on what you need.