In this design, I have assumed the existence of a working, RCF-compatible Natural Language Processing unit. Here, I will attempt to explain the basics of what is required of such a system to be of value to GHuRU.

The NLP unit must be able to convert most english text into RCF, and be able to make reasonable guesses regarding unknown words and phrases. A feedback learning loop could help GHuRU understand what made a good guess. Presumably some sort of learning algorithm would be used in this case, in conjunction with the Reliability Weighting system (the way by which the accuracy or the believability of some given information is determined).

GHuRU's modular design allows for parallel development of the different parts. This is where the necessity for the standard form (RCF) comes out. It allows different groups to work independently on the various modules of the system. As long as their modules adhere to the RCF specification, they will all work together.