Browse Prior Art Database

How to get a Large Natural-language System into a Personal Computer

IP.com Disclosure Number: IPCOM000127950D
Original Publication Date: 1986-Dec-31
Included in the Prior Art Database: 2005-Sep-14
Document File: 8 page(s) / 34K

Publishing Venue

Software Patent Institute

Related People

Bozena H. Thompson: AUTHOR [+4]

Abstract

The answer to the question of how to get a large natural-language system into a personal computer lies in the paging architecture of the system. The key is to use the input sentence, in conjunction with the lexicon and grammar table, to identify the minimal segments of both object code and data that must be brought into main memory. Once such a maximally paged architecture has been effectively imple-mented, it has wide ranging implications for process integration, networking and knowledge base distribution, and for the software engineering environment. The Natural Access System optimizes this architecture and exploits these implications. How to Get a Large Natural-Language System into a Personal Computer 77 The Natural Access System is a large natural-language sys-tem. It is now implemented and running on a personal com-puter (PC). This paper tells how we were able to get such a large system on such a little computer.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 11% of the total text.

Page 1 of 8

THIS DOCUMENT IS AN APPROXIMATE REPRESENTATION OF THE ORIGINAL.

How to get a Large Natural-language System into a Personal Computer

Bozena H. Thompson and Frederick B. Thompson

Computer Science Department California Institute of Technology

5215:TR:8C

Presented at the National Computer Conference, Chicago, July, How to get a large natural- language system into a personal computer

by BOZENA HENISZ THOMPSON and FREDERICK B. THOMPSON California Institute of Technology Pasadena, California

ABSTRACT

The answer to the question of how to get a large natural-language system into a personal computer lies in the paging architecture of the system. The key is to use the input sentence, in conjunction with the lexicon and grammar table, to identify the minimal segments of both object code and data that must be brought into main memory. Once such a maximally paged architecture has been effectively imple-mented, it has wide ranging implications for process integration, networking and knowledge base distribution, and for the software engineering environment. The Natural Access System optimizes this architecture and exploits these implications. How to Get a Large Natural-Language System into a Personal Computer 77

The Natural Access System is a large natural-language sys-tem. It is now implemented and running on a personal com-puter (PC). This paper tells how we were able to get such a large system on such a little computer.

THE OBJECTIVES OF NATURAL ACCESS

The Natural Access System has evolved over a number of years with the object of providing truly natural access to the computer for intelligent people who may not be programmers or even computer literate. Because of the academic setting of our research, we have been able to take a fresh look at this problem, and to guide our research through considerable experimentation.'-3 We have assumed that most people will have direct access to computers in civil and business organiza-tions, in research labs, on engineering floors, among manage-ment staffs, and in their homes. The forms and facilities of this access are still evolving in response to the question, what is the proper farm for this access and what are the requirements for a computer software system that will provide natural access to computers? Experience with existing systems makes it abundantly clear that almost every application of computers is a special applica-tion not adequately served by any single general purpose sys-tem, and that the requirements of any specific application are constantly in flux, thereby necessitating constant updating and extending of any system implemented for its support. Yet, it is not economically viable to program a new system from scratch for each application and to constantly retrofit that detailed special purpose design. What is more tenable is a system design that serves a defined range of applications, and within this range, is readily adaptable to each specific applica-tion; a system which in its

Californi...