Key findings from empirical studies—early design is opportunistic; critical role of pictures in design conception; impact of various cognitive limitations—have very effectively determined requirements and design for a set of tools to support early design. Key design features of the tools include respectively: (1) The (simultaneous) display of any software modules at arbitrary levels of abstraction and from any subsystems. The unrestricted, smooth navigation between these software modules. (2) Multiple design notations—pictorial and symbolic— cross-referenced, editable, and maintained consistent across all views. Integrated views of control flow, data flow, and functional decomposition. (3) Automatic layout at arbitrary levels of nesting. Visual display of execution paths in the solution. Automatic completeness and consistency check. Automatic visual indication and listing of modules with constraint violations.

In this descriptive and exploratory study, 32 users type help requests to what they believe is a computerized advisor. In fact, the advisor is a human mimicking realistic levels of intelligence and knowledge that can be expected from a computerized advisor. Results show that users request help with a very simple and restricted language that is characteristic of language generated under real-time production constraints and of child language. Moreover, users' utterances are frequently ungrammatical. It is hypothesized that these features arise from factors intrinsic to typed advisory situations: Users are performing a primary task under real-time constraints, and typing help requests is a secondary task. On the other hand, users refer to objects and events with very precise descriptions instead of faster-to-type pronouns; they produce very few ellipses and deictic expressions. Future research should elucidate whether shared context between users and computerized advisors needs to be richer than created in this study to sustain the use of expressions whose interpretations depend on context. The tuning of natural language interfaces to the features observed in this study may increase the usefulness of natural language interfaces to advisory systems. The presented methodology is a promising tool for further studies of these factors on users' language.

This study shows that top-down decomposition is problematic in the early stages of design. Instead, an opportunistic decomposition is better suited to handle the ill-structuredness of design problems. Designers are observed interleaving decisions at various levels of abstraction in the solution decomposition. The verbal protocols of three professionals designing a software system of realistic complexity are analyzed to determine the frequency and causes of opportunistic decompositions. The sudden discovery of new requirements and partial solutions triggered by data-driven rules and associations, the immediate development of solutions for newly discovered requirements, and drifting through partial solutions are shown to be important causes of opportunistic design. A top-down decomposition appears to be a special case for well-structured problems when the designer already knows the correct decomposition. Two cognitive models are briefly discussed in relation to opportunistic design. Finally, implications for training, methods, and computational environments to support the early stages of design are outlined.

Advisory systems can be very powerful general tools for users. Formal query languages, menus, and direct manipulation interfaces might not suffice to access advisory systems' full functionality. The capabilities of natural language interfaces could be required. Unfortunately, natural language interfaces are not meeting the needs yet. Wide syntactic coverage is often traded off against handling ungrammatical sentences. However, this study shows that users request help with a very simple and restricted English, characteristic of unplanned or of child language. Moreover, users' utterances are frequently ungrammatical. It is argued that the simple syntax and the ungrammaticalities are determined by features intrinsic to advisory systems: users request help by typing to perform another primary task under real-time production constraints. Because of intrinsic performance constraints, users naturally resort to earlier and simpler forms of syntax. Natural language interfaces to advisory systems need not cover a wide variety of syntactic constructions but they must emphasize robust parsing.

A verbal protocol study of professional software designers has revealed three design process control strategies. At least one of them, the generation of opportunistic solutions at different levels of detail accompanied by problem domain modeling, had not been observed in previous empirical studies nor had been acknowledged in the software engineering practices. Specific breakdowns (difficulties) were associated with the different design process control strategies. Software tools should be provided to designers to alleviate these breakdowns. Parts of a cognitive model of software design, based on distributed control from specialists such as design schemas, design heuristics, and design methods, are presented to account for the observed control strategies.

Article Free Access Share on Video: Data for studying human-computer interaction Authors: W. E. Mackay Massachusetts Istitute of Technology Massachusetts Istitute of TechnologyView Profile , R. Guindon Microelectronics and Computer Technology Corporation Microelectronics and Computer Technology CorporationView Profile , M. M. Mantel Electric Data Systems Corporation Electric Data Systems CorporationView Profile , L. Suchman Xerox Palo Alto Research Center Xerox Palo Alto Research CenterView Profile , D. G. Tatar Xerox Palo Alto Research Center Xerox Palo Alto Research CenterView Profile Authors Info & Claims CHI '88: Proceedings of the SIGCHI Conference on Human Factors in Computing SystemsMay 1988 Pages 133–137https://doi.org/10.1145/57167.57189Published:01 May 1988Publication History 15citation544DownloadsMetricsTotal Citations15Total Downloads544Last 12 Months56Last 6 weeks6 Get Citation AlertsNew Citation Alert added!This alert has been successfully added and will be sent to:You will be notified whenever a record that you have chosen has been cited.To manage your alert preferences, click on the button below.Manage my AlertsNew Citation Alert!Please log in to your account Save to BinderSave to BinderCreate a New BinderNameCancelCreateExport CitationPublisher SiteeReaderPDF

User-adviser dialogues were collected in a typed Wizard-of-Oz study ("man-behind-the-curtain study"). Thirty-two users had to solve simple statistics problems using an unfamiliar statistical package. Users received help on how to use the statistical package by typing utterances to what they believed was a computerized adviser. The observed limited set of users' grammatical and ungrammatical forms demonstrates the sufficiency of a very restricted grammar of English for a natural language interface to an advisory system. The users' language shares many features of spoken face-to-face language or of language generated under real-time production constraints (i.e., very simple forms of utterances). Yet, users also appeared to believe that the natural language interface could not handle fragmentary or informal language and users planned or edited their language to be more like formal written language (i.e., very infrequent fragments and phaties). Finally, users also appeared to believe in poor shared context between users and computerized advisers and referred to objects and events using complex nominals instead of faster-to-type pronouns.

Novice users engaged in task-oriented dialogues with an adviser to learn how to use an unfamiliar statistical package. The users' task was analyzed and a task structure was derived. The task structure was used to segment the dialogue into subdialogues associated with the subtasks of the overall task. The representation of the dialogue structure into a hierarchy of subdialogues, partly corresponding to the task structure, was validated by three converging analyses. First, the distribution of non-pronominal noun phrases and the distribution of pronominal noun phrases exhibited a pattern consistent with the derived dialogue structure. Non-pronominal noun phrases occurred more frequently at the beginning of subdialogues than later, as can be expected since one of their functions is to indicate topic shifts. On the other hand, pronominal noun phrases occurred less frequently in the first sentence of the subdialogues than in the following sentences of the subdialogues, as can be expected since they are used to indicate topic continuity. Second, the distributions of the antecedents of pronominal noun phrases and of non-pronominal noun phrases showed a pattern consistent with the derived dialogue structure. Finally, distinctive clue words and phrases were found reliably at the boundaries of subdialogues with different functions.

Anaphora resolution is the process of determining the referent of anaphors, such as definite noun phrases and pronouns, in a discourse. Computational linguists, in modeling the process of anaphora resolution. have proposed the notion of focusing. Focusing is the process, engaged in by a reader of selecting a subset of the discourse items and making them highly available for further computations. This paper provides a cognitive basis for anaphora resolution and focusing. Human memory is divided into a short-term, an operating, and a long-term memory. Short-term memory can only contain a small number of meaning units and its retrieval time is fast. Short-term memory is divided into a cache and a buffer. The cache contains a subset of meaning units expressed in the previous sentences and the buffer holds a representation of the incoming sentence. Focusing is realized in the cache that contains a subset of the most topical units and a subset of the most recent units in the text. The information stored in the cache is used to integrate the incoming sentence with the preceding discourse. Pronouns should be used to refer to units in focus. Operating memory contains a very large number of units but its retrieval time is slow. It contains the previous text units that are not in the cache. It comprises the text units not in focus. Definite noun phrases should be used to refer to units not in focus. Two empirical studies are described that demonstrate the cognitive basis for focusing, the use of definite noun pphrases to refer to antecedents not in focus. and the use of pronouns to refer to antecedents in focus.