Abstract A search for information can be viewed as a series of decisions made by the searcher. Two dimensions of the search environment affect a user's decisions: the user's knowledge, and the configuration of the information retrieval system. Drawing on previous findings on users' lack of search or domain knowledge, this article investigates what the user needs to know to make informed search decisions at the United States Bureau of Labor Statistics (BLS) Web site, which provides statistical information on labor and related topics. Its extensive Web site is a rich collection of statistical information, ranging from individual statistics such as the current Consumer Price Index (CPI), to a large statistical database called LABSTAT that can be queried to construct a table or time series on the fly. Two models of the search environment and the query process in LABSTAT are presented. They provide complementary views of the decision points at which help may be needed, and also suggest useful help content. Extensive examples based on the industry concept illustrate how the information could assist users' search decisions. The article concludes with a discussion of the role of help facilities in Web searching, and the interesting question of how to initiate the provision of help.

Previous research describing Web page and link classification systems resulting from a content analysis of over 75 Web pages left us with four unanswered questions. 1 What is the most useful application of page types: as descriptions of entire pages or as components that are combined to create pages? 2 Is there a kind of analysis that we can perform on isolated anchors, which can be text, icons, or both together, that is equivalent to the syntactic analysis for embedded and labeled anchors? 3 How explicitly are readers informed about what can be found by traversing a link, especially for the relatively broad categories of expansion and resource links? 4 Is there a relationship between the type of link and whether its target is a whole page or a fragment, or if its target is in the same site or a different site than its source? This article examines these questions under the assumption that the author and the reader of Web pages will cooperate in order to have successful communication. Our discussion leads to ideas of how author-provided context and readers' expectations and experience are combining to form new stylistic conventions and genres on the Web.

The research presented here describes a method for automatically identifying sublanguage (SL) domain terms and revealing the patterns in which they occur in text. By applying this method to abstracts from a variety of disciplines, differences in how SL domain terminology occurs can be discerned. Results of this research have both practical and theoretical implications. These include 1) the identification of patterns of domain term occurrence, 2) a step toward the identification of families of SLs that share term occurrence patterns, 3) a domain term extraction procedure that can exploit the term occurrence patterns, and 4) evidence to support the intuitive notion of a continuum of “technicality” of disciplines and their SLs. The investigation revealed relatively consistent differences between the hard sciences, such as physics or biology, and the social sciences and humanities, such as history or sociology. The hard sciences tended to have more domain terms, and more of these terms occurred in sequences than in the social sciences and humanities. The seed terms used in this research occurred adjacent to domain terms more often in the hard sciences than in the social sciences. The extraction process was more successful in the hard science disciplines than in the social sciences, identifying more of the domain terms while extracting fewer general terms.

The use of terms from natural and social scientific titles and abstracts is studied from the perspective of sublanguages and their specialized dictionaries. Different notions of sublanguage distinctiveness are explored. Objective methods for separating hard and soft sciences are suggested based on measures of sublanguage use, dictionary characteristics, and sublanguage distinctiveness. Abstracts were automatically classified with a high degree of accuracy by using a formula that considers the degree of uniqueness of terms in each sublanguage. This may prove useful for text filtering or information retrieval systems. © 1995 John Wiley & Sons, Inc.

Quotation provides a link between the work in which it is used, and the work from which it was taken. The research described in this article extends work done by Kilgour and Feder (1992) on the frequency of direct quotations, looking at the use of incomplete sentence quotations (ISQs). Specifically, this research attempts to determine if (a) there are any patterns of usage in the parts of speech that are quoted in ISQs; (b) there are any patterns of usage in how the ISQs are incorporated into text; (c) there are any systematic differences in ISQ usage between four disciplinary areas; and (d) there are any systematic differences in ISQ usage between scholarly books and journals. Results showed no major differences in the numbers of ISQs in the different areas in books, but the journals formed two distinct groups, with the Arts and Humanities journals having many more ISQs than those in Social Science and Science and Technology. Nevertheless, the distribution of ISQs into the parts of speech was remarkably constant, with NPs and VPs making up about three quarters of the ISQs. Very few ISQs were marked with explicit cue words. © 1993 John Wiley & Sons, Inc.

Article Free Access Share on The constituent object parser: syntactic structure matching for information retrieval Authors: D. P. Metzler Department of Information Science, University of Pittsburgh Department of Information Science, University of PittsburghView Profile , S. W. Haas Department of Information Science, University of Pittsburgh Department of Information Science, University of PittsburghView Profile Authors Info & Claims SIGIR '89: Proceedings of the 12th annual international ACM SIGIR conference on Research and development in information retrievalMay 1989Pages 117–126https://doi.org/10.1145/75334.75348Published:01 May 1989Publication History 6citation344DownloadsMetricsTotal Citations6Total Downloads344Last 12 Months15Last 6 weeks1 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

The architecture and functioning of the Constituent Object Parser are described. This system has been developed specifically for text processing applications such as information retrieval, which can benefit from structural comparisons between elements of text such as a query and a potentially relevant abstract. The general way in which the system performs these matches, and the ways in which this objective influenced the design of the system are described. The parsing architecture incorporates several interesting features including: (1) an unusual combination of declarative and procedural representation techniques, (2) a monotonic discipline which permits useful heuristic approaches to difficult linguistic problems such as ellipsis, conjunction, ambiguity, and ill-formed and incomplete input, and (3) an attempt to minimize the level of syntactic detail required in both the grammar and the lexicon. © 1989 John Wiley & Sons, Inc.

The Constituent Object Parser is a shallow syntactic parser designed to produce dependency tree representations of syntactic structure that can be used to specify the intended meanings of a sentence more precisely than can the key terms of the sentence alone. It is intended to improve the precision/recall performance of information retrieval and similar text processing applications by providing more powerful matching procedures. The dependency tree representation and the relationship between the intended use of this parser and its design is described, and several problems concerning the processing of ambiguous structures are discussed.