As I understand it, each entry in the database contains all of the tags shown in the details screen, and the database size is related to the track count limit and search speed...

So why does the the database contain Comments, Format, File Size, none of which are useful search parameters.

Wouldn't it be better to only database the tags needed to support search etc, and to pull the other information directly from the mp3 file when the empeg starts to cache it? We'd gain efficiency by having a smaller database, and we'd also be in a position to remove the arbitary comment length.