This paper describes a natural language processing system implemented at Hewlett-Packard's Computer Research Center. The system's main components are: a Generalized Phrase Structure Grammar (GPSG); a top-down parser; a logic transducer that outputs a first-order logical representation; and a "disambiguator" that uses sortal information to convert "normal-form" first-order logical expressions into the query language for HIRE, a relational database hosted in the SPHERE system. We argue that theoretical developments in GPSG syntax and in Montague semantics have specific advantages to bring to this domain of computational linguistics. The syntax and semantics of the system are totally domain-independent, and thus, in principle, highly portable. We discuss the prospects for extending domain-independence to the lexical semantics as well, and thus to the logical semantic representations.

The complexity of information processing, disseminating and controlling is very high within a sizable corporation. Database design, targeted to carry out these functions, is constantly improving. Nevertheless, nonprogrammers have trouble accessing most databases. Layers of EDP personnel, as well as unresponsive and cumbersome systems, are an obstacle to effective database use.In this paper, we describe an exploratory investigation of databases based on REL ENGLISH [1]. In our system, data structures and access language were designed to map closely the corporate structure and its terminology. Our test vehicle was a portion of Hewlett-Packard's internal information network, the corporate Order Processing System. We have identified about a dozen job related perspectives belonging to geographically dispersed and functionally stratified end-users within various entities of the HP organization. A multinested data model is an intertwined hierarchy: the join of separate hierarchies with different lexicons but shared data events. For our test, we have selected two major intersecting multinested user views: sales and manufacturing. Our design accommodates users with differing perspectives of this model.The system data structures were constructed using the REL ENGLISH primitives. The Condensed Order Records (COR) base spanned a multinested hierarchy four levels deep for the Sales Organization taxonomy and five levels deep for the Manufactured Products taxomony. The design permits the user to query the COR base about the data model itself. He can, of course, also obtain the standard statistical views of the data.Corporate taxonomies are dynamically changing structures. The database model needs to reflect this change. In most cases, the user wants to manipulate data at three levels in his own job taxonomy and at all levels in the other taxonomies. The multinested data model is needed in order to permit the user to view the data from the same perspective that he views his job in the corporate organization.