Radarbox uses SQLite databases, and the journals should be getting commited on a regular basis. If you find that the journal files are getting big, then it would indicate that the journals aren't getting commited by Radarbox. I've never seen this happen, but others have and it results in data being lost from Mylog, for example.
This needs a more definitive answer from Airnav, but I've always had concerns that under some conditions database updates remain in the journals, and then get lost. Possibly a failure in atomic commits in Radarbox? Or maybe an interation between the NTFS journal system and SQLite journal system. I'w always put SQLite databases on a FAT32 drive to avoid NTFS journalling - it's an overhead you just don't need.
Dave, you might want to post an explanation on the other forum. I also tried answering this on the other forum before being banned again for being helpful :(