Mork is a more or less sick database format for text files . Some of the ideas behind it are good, some are ... bad. Nevertheless, Mork is used in some widespread applications and up to today, there is no independent library available which can handle Mork files.
This project's goal is to provide a Java implementation for the Mork format. The primary goal of jMork is to provide read functionality to database files. The focus is not on writing Mork files. The Mozilla applications which use Mork files (Firefox, Thunderbird) will replace the format in the future nevertheless.
Mork consists of dictionaries (key/value maps), database items like table, row, cell and a transactional-style semantic called groups. The data can be stored as literal values or as referenced identifiers, which point to values in the dictionaries. The data itself is contained in tables or loose rows. Cells are simple assignments (key=value).
Some effort has been put into how values are escaped, how compression could be made more effective. But these efforts are not worth the trouble which you have parsing the format.