It would be convenient to auto-detect file formats. This is a hard problem, and one where the other Bio* projects differ. BioPerl has pretty good support for guessing file formats based on extension, magic bytes, or contents. BioPython has adhered to the python philosophy of 'explicit is better than implicit', and required the user to provide the format type.

What's the BioJava philosophy? I ask because I would like to implement auto-detecting structure formats (e.g. PDB vs mmCIF vs XML).