A NATURAL LANGUAGE INTERFACE FOR DISTRIBUTED HETEROGENEOUS DATABASES FOR AIR FORCE APPLICATIONS

TO AUTOMATICALLY RESOLVE AND PROCESS DATA TO APPLICATIONS REGARDLESS OF THE SOURCE REQUIRES THE USE OF MULTIPLE DISPARATE DATABASES AND DISTRIBUTED INFORMATION SYSTEMS. TO STORE AND EXTRACT INFORMATION FROM SUCH SOURCES, INCLUDING RELATIONAL, HIERARCHIAL AND DISTRIBUTED DATABASES AS WELL AS OTHER SOURCE (I.E. TEXT FILES, HYPERMEDIA, SPREADSHEETS), USERS ARE FORCED TO LEARN A PLETHORA OF DIFFERENT LANGUAGES AND ACCESS METHODS. TO ALLEVIATE THIS DIFFICULTY, WE PROPOSE A COMMON TRANSPORTABLE NATURAL LANGUAGE INTERFACE (TNLI) WHICH WILL TRANSLATE ENGLISH LANGUAGE USER REQUESTS TO STORE OR RETRIEVE RELAVANT INFORMATION VIA NETWORKING TO OR FROM ALL THE DATABASES. THE OBJECTIVE IS TO MAKE AVAILABLE LARGE VOLUMES OF DISPARATE INFORMATION IN A USER-FRIENDLY MANNER. THIS APPROACH RENDERS THE ENTIRE INFORMATION SYSTEM A "VIRTUAL COMMON MEMORY DATA PROCESSOR"--THERE IS ABSOLUTELY NO NEED FOR A CUSTOM MACHINE TO BE BUILT. A TRANSPORTABLE FRONT END INDEPENDENT OF THE DATABASES WILL PRODUCE LOGICAL FORM FRAGMENTS (LFFS) THAT ARE FED TO A PLANNER. THEN THE PLANNER USES KNOWLEDGE OF THE DATABASE SEMANTICS AND THE LFF TO DIRECT QUERIES TO EACH APPLICABLE DATABASE.