To complete design of the system, in addition to the nametable implementation and search function, components are needed for deletion of records. We will not pursue this topic here, but clearly the issues of storage reclamation and allocation would have to be considered.
Beyond this, there are some more basic unresolved problems. Exactly how are questions asked of the system?
One way is for anyone wanting answers to questions to bring them to the system's designers. The designers then write programs to output the solutions. This is feasible in certain situations; in fact, many organizations do information retrieval in this way. A related solution would be to give the data representation to questioners and have them write their own programs.
Another way is to design a simple "menu," or list, of questions that includes at least those most likely to be asked. Users could then "check off" their questions if they appear on the menu, and previously written programs could be run to answer them. This solution proves satisfactory as long as the questions people want answered are on the menu.
A more sophisticated scheme would be to design a general query language. Questioners could then ask their questions in this language. A built-in language translator would interpret the question and automatically produce a program to print the answer. Achieving this solution can be quite complex and might challenge the state of the art at the present time.
The issue of security has been ignored. It would be of major concern in any retrieval system. Users' access to specific information and ability to modify it would have to be limited.