On the other hand the standalone models would all need "basic" knowledge, so I'm assuming this would be much less disk space-efficient.