> I have directories on my filesystem(s), files in those directories, and amazingly powerful applications to navigate them (Total Commander / mc / Krusader).I think that the problem is that a file system typically represents a single type of hierarchy. Maybe you have your books in the file system arranged by topic->author->book, but other arrangements like topic->year->book, book->year (for editions) or author->book->author->book->... (for citations) can be useful. Maybe you want to list books that relate to a set of multiple topics written by either of two specific authors.
In Unix you can solve this by using symbolic links, which may become tedious to maintain (but less so by using automated tools). Generally, you can sort of address it by including meta information in the file names (like the warez scene does), but that is tedious when you have a lot of potentially relevant meta information.
The benefits of throwing all the information into a relational database and querying into that instead of a regular PC file system is that you can define the desired hierarchy and what meta-data is important on the fly and effectively create a file system (in the pre-computer sense) and perform a search at the same time.
I haven't looked closely enough at this software to say that its "library" actually addresses these concerns, but for large libraries there are obvious benefits to 1) Using multiple hierarchies or faceted classification for information query 2) defining such hierarchies or sets of classifying categories on the fly. These aren't new solutions, and fundamentally the organization of information in a library is not a new problem, so I wouldn't call your position old school.