Using a natural language interface with casual users

doi:10.1016/S0020-7373(08)80007-7

International Journal of Man-Machine Studies

Volume 32, Issue 3, March 1990, Pages 341-361

https://doi.org/10.1016/S0020-7373(08)80007-7 Get rights and content

Abstract

Although there is much controversy about the merits of natural language interfaces, little empirical research has been conducted on the use of natural language interfaces for database access, especially for casual users. In this work casual users were observed while interacting with a real-life database using a natural language interface, Intellect.

Results show that natural language is an efficient and powerful means for expressing requests. This is especially true for users with a good knowledge of the database contents regardless of training or previous experience with computers. Users generally have a positive attitude towards natural language. The majority of errors users make are directly related to restrictions in the vocabulary. However, feedback helps users understand the language limitations and learn how to avoid or recover from errors. Natural language processing technology is developed enough to handle the limited domain of discourse associated with a database; it is simple enough to support casual users with a general knowledge of the database contents; and it is flexible enough to assist problem-solving behaviour.

References (27)

A.W. Biermann et al.
An experimental study of natural language programming
International Journal of Man-Machine Studies
(1983)
M.J. Kelly et al.
Limited vocabulary natural language dialogue
International Journal of Man-Machine Studies
(1977)
K. Morik
Customers' requirements for natural language systems: results of an inquiry
International Journal of Man-Machine Studies
(1984)
Artificial Intelligence Corp
(1986)
J.M. Carroll et al.
Interface design issues for advice-giving expert systems
Communications of the ACM
(1987)
W.R. Ford et al.
Self limited and unlimited word usage during problem solving in two communication modes
Journal of Psycholinguistics Research
(1979)
L. Harris
The ROBOT system: natural language processing applied to data base query
A.G. Hauptman et al.
A comparison of command, menu-selection and natural language computer programs
Behaviour and Information Technology
(1981)
G.G. Hendrix
Natural-language interface
American Journal of Computational Linguistics
(1982)
G.W. Hopkins
Natural language database access: a field trial of intellect

M. Jarke et al.

A field evaluation of natural language for data retrieval

IEEE Transactions of Software Engineering SE-II

(1985)

M. Jarke et al.

A framework for choosing a database query language

ACM Computing Surveys

(1985)

J.F. Kelley

An empirical methodology for writing user-friendly natural-language computer applications

Cited by (19)

Programming language, natural language? Supporting the diverse computational activities of novice programmers
2017, Journal of Visual Languages and Computing
Citation Excerpt :
Of the English sentences typed, 81% could be interpreted, and the authors suggest that many of the errors that did occur were easily resolvable. Capindale and Crawford [6] also found that natural language is an effective means of expressing database queries, particularly for users with prior knowledge of the contents of the database, regardless of their aptitude with computers. Miller carried out a number of studies investigating the potential for natural language to be used for programming [31–33].
Given the current focus on teaching computational concepts to all from an early age, combined with the growing trend to empower end users to become producers of technology rather than mere consumers, we consider the issue of “computational notation”. Specifically, where the goal is to help individuals develop their understanding of computation and/or use computation in real world settings, we question whether natural language might be a preferred notation to traditional programming languages, given its familiarity and ubiquity. We describe three empirical studies investigating the use of natural language for computation in which we found that although natural language provides support for understanding computational concepts, it introduces additional difficulties when used for coding. We distilled our findings into a set of design guidelines for novice programming environments which consider the ways in which different notations, including natural language, can best support the various activities that comprise programming. These guidelines were embodied in Flip, a bi-modal programming language used in conjunction with the Electron toolset, which allows young people to create their own commercial quality, narrative based role- playing games. Two empirical studies on the use of Flip in three different real world contexts considered the extent to which the design guidelines support ease of use and an understanding of computation. The guidelines have potential to be of use both in analysing the use of natural language in existing novice programming environments, and in the design of new ones.
The impacts on user performance and satisfaction of a voice-based front-end interface for a standard software tool
1996, International Journal of Human Computer Studies
The objective of this study was to compare empirically the effect on performance and satisfaction of a menu and a front-end voice interface to a commonly used spreadsheet software package. In this study, the type of human-computer interface used (standard keyboard/mouse use of menus or keyboard/mouse with a voice front-end) is expected to influence user performance (task completion time and error rates) and satisfaction. The user's novice/expert classification is expected to interact with these two types of interfaces to influence the efficiency of the user's performance. The results suggest that there are significant relationships between task performance and level of expertise/type of interface and between user attitudes and type of interface. In general, the front-end voice interface users performed worse and had less favorable attitudes towards the software tool than the menu interface users.
A semantic net representation for the classification of minerals
1995, Computers and Geosciences
A program is described which uses a variation of a semantic net, the tree network, to represent the classification of minerals together with their properties and values. The tree network efficiently represents the hierarchical nature of the mineral classification. The inheritance and recognition principles characteristic of the semantic network underlie the query mechanism. PROLOG is used as the representation language, this provides a declarative method of representing the tree network and allows queries to be posed in a natural user-friendly manner. Minerals are represented by nodes in the tree network which are connected by the structural links IS-A and ARE. Properties and their values are connected to mineral nodes by (attribute, property) and (property, value) pairs. Several different search strategies are described for performing recognition and inheritance type queries of the tree network.
Effects of linguistic sophistication on the usability of a natural language interface
1993, Interacting with Computers
Casual users of database systems are not skilled in structured query languages. Natural language interfaces (NLIs) could provide them with the flexibility and ease of use that they require. The paper challenges the dictum that the usability of an NLI is enhanced when its linguistic capabilities are extended. It is argued that effective natural language communication needs a naturalistic sublanguage of English, reduced in complexity, but nevertheless providing the flexibility of natural language input. Two investigations are described, both of which involved real users performing real information retrieval tasks. The first gives an insight into the detailed characteristics of such a sublanguage, providing a comparison with earlier research. The second compares the effect of intersentential linking devices like ellipsis, on the usability of an NLI, with that of simple, extralinguistic editing facilities. The results show that enhanced linguistic capabilities can indeed improve usability under certain circumstances, but that extralinguistic enhancements can be just as effective. The results also show that usability can actually degrade as both the linguistic and the extralinguistic capabilities of an interface are improved.
Accomodating domain-independence-a new approach to the development of general natural language interfaces
1992, Engineering Applications of Artificial Intelligence
This paper proposes a domain-independent approach to constructing natural language interfaces for applications. Domain independence provides generality to natural language interfaces. A hierarchical structure for natural language processing and a representation language, the Intermediate Carrier Language, are introduced. A special phase of natural language processing called domain compiling is also proposed and described. To assess and improve this approach, an experimental natural language interface KZ2/TZJ has been developed. This serves as the common Chinese interface for both Unix and Vi. Unix and Vi were chosen as KZ2/TZJ's applications because they belong to different kinds of application categories and using them may give a fair evaluation of the approach. Experimental results indicate that this approach is a good candidate for developing general natural language interfaces.
An expert natural language interface for statistical packages
1992, Expert Systems With Applications

View all citing articles on Scopus

^†: This research was supported by a grant from the National Science and Engineering Research Council of Canada

View full text

Using a natural language interface with casual users†

Abstract

International Journal of Man-Machine Studies

International Journal of Man-Machine Studies

International Journal of Man-Machine Studies

Artificial Intelligence Corp

Interface design issues for advice-giving expert systems

Communications of the ACM

Self limited and unlimited word usage during problem solving in two communication modes

Journal of Psycholinguistics Research

The ROBOT system: natural language processing applied to data base query

A comparison of command, menu-selection and natural language computer programs

Behaviour and Information Technology

Natural-language interface

American Journal of Computational Linguistics

Natural language database access: a field trial of intellect

A field evaluation of natural language for data retrieval

IEEE Transactions of Software Engineering SE-II

A framework for choosing a database query language

ACM Computing Surveys

An empirical methodology for writing user-friendly natural-language computer applications