An uberdictionary in steno is a collection of translations with context about their pronunciations, orthography, origin, and other pertinent attributes.
With an uberdictionary, you can write context for translations once, and every single theory that uses the uberdictionary can use those translations without having to create them anew.
If you’re using a rule-based generated dictionary (via quartz or potentially others), you get new entries completely for free.
Theories should be able to choose what translations they support, and build rules around them without immense effort.
This is a very important concept in quartz.