Resource Library: Multimedia

home *** CD-ROM | disk | FTP | other *** search

/ Resource Library: Multimedia / Resource Library: Multimedia.iso / hypertxt / msdos / hypdiss / chapt1 < prev next >

Wrap

Text File | 1992-04-11 | 24.4 KB | 426 lines

CHAPTER I INTRODUCTION Introduction Hypertext software applications seem to promise delivery of information systems providing both easy and efficient information retrieval. A hypertext system is a computerized network of information nodes or text segments connected by means of visible reference pointers, or relational "links" (Conklin 1987) <refs 6 1>. These links can also execute operating system commands or interfaces to external systems. This would include calls to external programs, to remote databases, and to Compact Disk-Read Only Memory (CD- ROM), video, or audio devices. In a hypertext system, the user moves a display screen pointer or highlight to select hypertext links. The person uses cursor arrow keys or a mouse pointing device to move the pointer. A visible indicator or highlight on the display screen gives visual feedback while "pointing" at links. Finally, a single keystroke or mouse click activates the link. Activation of a link results in an instant "jump" or display of the corresponding document or text node, or execution of the associated command sequence. The combination of linked information resources or associated sequences of actions means that a hypertext system is not simply a software type. It is rather an integration or mixture of computer hardware and software, designed for the presentation of information to users, which allows them to easily manage the information that it contains (National Legal Research Systems 1990) <refs 18 6>. The user is not restricted to using the text or multimedia information in an author's prescribed pattern. Instead, users may move or "jump" around at will, selecting the pathways or information approaches that appear most suitable or rewarding (Seyer 1991) <refs 20 9>. The origin of the hypertext concept is generally credited to Vannevar Bush, whose 1945 Atlantic Monthly article introduced the idea of associative links, as well as many other seminal information retrieval concepts. Bush envisioned a fictional personal information system, which he called "Memex." The imaginary system used personally assigned associative links, to retrieve material containing the user-defined information or ideas (Bush 1945) <refs 4 10>. This early system concept was based upon the limited mechanical retrieval and photographic image technology of the World War II era. Nevertheless, Bush was able to formulate and describe the link retrieval idea. He compared it to the natural operation of the human mind. He contrasted this mental associative linking mechanism, operating by intuitive connection of one idea or concept to another, with traditional classified or hierarchical information retrieval systems. He viewed these as being much more difficult to use. As he put it: The human mind does not work that way. It operates by association. With one item in its grasp, it snaps instantly to the next that is suggested by the association of thoughts, in accordance with some intricate web of trails carried by the cells of the brain. . . . [T]he speed of action, the intricacy of trails, the detail of mental pictures, is awe-inspiring beyond all else in nature. (Bush 1945, 106) <refs 4 10> Bush described the system as using vacuum tube-level electronics. He also included electromechanical units, and microfilm data storage devices. The idea was not realistically executable, since it was limited by the technology of the time. However, developers working with more advanced technology were able to produce working hypertext systems in the 1960s. In 1968, Douglas Engelbart produced the NLS/Augment system (Engelbart 1988) <refs 8 12>. Coincidentally, in efforts to produce more usable computer systems, Engelbart and his group originated many of the innovations giving rise to the modern concepts of office automation and the personal computer. Among his contributions were the mouse pointing device, text processing, the first major implementations of electronic mail and online group conferencing, outlining software, and the multiple windowing concept (Horn 1989) <refs 14 6>. Later, at Xerox PARC Labs (Palo Alto Research Center), for the NoteCards system, investigators combined high-definition graphic displays and the mouse pointing device into a personal system workstation operating on an early desktop computer (Halasz 1988) <refs -halasz>. Ted Nelson, another early hypertext developer, coined the term "hypertext" in 1965, from a combination of the Latin roots "hyper" and "text." In explaining his usage, he described "hyper" as "extended, generalized, and multidimensional" (Nelson 1973, p. M22-M23) <refs -nelson>. Hypertext technology finally became practical in the late 1980s. The development was stimulated by the concurrent availability of substantial personal computing power. This resulted from the economical disk storage, cheap random access memory (RAM), and the development of relatively inexpensive high-quality displays. The personal systems resulted from the combination of technological capacity and economy. This contrasted with the previous restriction of resources to institutional computing purposes (Locatis 1987 <refs -locatis>; Nielsen 1990 <refs 18 20>). The increasing availability of powerful, responsive software packages, designed for easy use, has resulted in substantial growth in the number of computer application programs. Ease of use and "user-friendliness" has been a major factor in this increase. Hypertext systems are quite attractive to "end-users" of automated information systems, because of their simplicity of operation. With hypertext, there is no need to learn complicated retrieval command languages, nor to carefully and precisely use proper syntax, nor to endure constant, incommunicative, error messages. Instead, links or visible pointers representing associated information or command sequences can easily be selected. This is generally done by simply "pointing" at or highlighting the visibly displayed depictions using cursor keys or a mouse device. A single keystroke or mouse click can then activate the selected link. This direct execution allows for quick and intuitive use of an otherwise sophisticated system. This "point & shoot" operating metaphor, using simple eye/hand coordination operations for execution of complex information system commands, has been described as direct manipulation (Shneiderman 1987a <refs 20 24>). The use of direct manipulation techniques in the computer-human- interface is characteristic of recent, more sophisticated computer system design principles. Researchers find that execution of complex computer processes by simple or direct user action results in the greatest leverage of human abilities (Baecker and Buxton 1987 <refs -baecker>). A simple explanation for the effectiveness of the direct manipulation approach is that users simply find them easier and more preferable to use. This appears to be an obvious example of the "principle of least effort," as expressed by Zipf (1972) <refs -zipf>. This inclination to expend minimal effort appears to be a natural human preference, evident in such diverse examples as word processing, photocopying, automobile automatic transmissions, or television remote controls. Personal computer users have exhibited a strong marketplace preference for the ease of use, and minimum effort. They are similarly gravitating towards use of hypertext applications, at present found mainly in online help modules. Bush's associative linking hypertext approach has not been limited to information retrieval applications. The hypertext metaphor has already served as an efficient approach to simplifying the general operation of computer systems. The point & shoot hypertext interface has gained widespread end-user acceptance, evidenced by the success of the Macintosh and Windows operating systems in the late 1980s and early 1990s. The hypertext concept of the visual links and associated actions or commands is the dominant metaphor for the "Graphic User Interface" (GUI) of these two systems. In the graphic interface, diagrammatic "icons," are associated with operating system or application program commands. This visual association metaphor or logic is functionally equivalent to the logic underlying hypertext links. Hypertext links have become extremely popular in "help screen" online documentation, within application programs. In these systems, a user displays online help by simply pointing at a word or phrase using mouse or cursor keys, and then clicking or using hypertext function keystrokes to retrieve the related information. Many authors have emphasized the knowledge and information system integrating aspects of the hypertext approach (Beck and Spicer 1988 <refs -spicer>; Chignell and Lacy 1988 <refs -chignell>; Jones 1990 <refs 15 5>; Smith 1988 <refs 21 21>; Underwood 1988 <refs -underwood>). They point out that the hypertext mechanism makes it easy to integrate or bind together knowledge or information stored on different computer systems, different hardware platforms, or even in different media formats. Locatis (1987) <refs -locatis> emphasizes this, pointing out that systems designers may design complex multimedia information or learning systems, mixing and matching supporting formats, yet retain complete ease of use. For example, the systems can interweave textual, graphic, audio, and video materials, as well as remote online systems, into a single coordinated multimedia system (Yankelovitch et al. 1988) <refs 24 22>. It is evident that hypertext is a promising software technology. It appears to have great potential for information retrieval applications. Ease of operation combined with information retrieval implementations would appear to have high value for reference publishing and information dissemination purposes. Statement of Purpose This will be an exploratory study addressing the information retrieval potentials of hypertext information retrieval systems. It will study the implementation of electronic analogs to traditional print-oriented information access techniques. The paper will present a detailed case study of an existing hypertext authoring system. It will identify traditional information retrieval or access methods used in the case study subject system. It will then recommend additional library and information science access models with potential for successful transfer to the hypertext medium. The study will conclude with generalization about the feasibility of transferring these formal information access mechanisms to the hypertext medium. Problem Statement Hypertext is an effective, "user-friendly," information delivery technology. Up to now, the details of technical implementation and user interface design have dominated published hypertext system research. A few papers have featured the planning or design of consistent information retrieval methods for hypertext systems (Egan et al. 1989a <refs 7 20>, 1989b <refs 8 1>; Frisse 1988b <refs 9 17>; Glushko 1988, 1989, 1990 <refs -glushko>). However, there has been no substantial study of information organizational and control methods applied to hypertext information system design. Study Rationale The study will address the transferability of traditional information organization and access models to automated hypertext information systems. Traditional library science methodologies such as classification, cataloging, controlled vocabularies, and indexing have functioned for providing information identification and filtering mechanisms. In the past, these information "filters" have proven effective for accessing printed literature and other forms of recorded information (Meadow 1973 <refs -meadow>). Using these traditional access tools, users have been able to locate material of probable interest, within a global mass of information. Traditional library science and archival methods stress standardization and control in these information access models (Fugmann 1985 <refs -fugmann>; Davis and Rush 1979 <refs -davis>). These library technical methods attempt to remedy problems inherent to language-based information retrieval systems. These difficulties are caused by such factors as inconsistent vocabulary, synonymy, homonymy, association, and level of hierarchy (Cleveland and Cleveland 1990 <refs 5 14>, Foskett 1982 <refs -foskett>, Meadow 1973 <refs -meadow>). Hypertext software takes full advantage of automated technology. It adds the capability of fast, yet effortless, display or delivery of desired information units. Hypertext offers a simple and effective direct manipulation interface, providing associative retrieval links to relevant or related information (Urr 1991 <refs 23 8>). Hypertext system authors can effectively incorporate or emulate traditional information access approaches by implementing them using links. Until recently, computer science and cognitive psychology researchers have dominated the hypertext literature. As a result, most research has addressed computer system implementations and user interface design. There has thus been little attention to information authoring methods and knowledge base design. The emphasis on implementations has resulted in hypertext systems which are efficient in retrieval and display technology, and highly "user-friendly" in operation. Unfortunately, they have also tended to be subjectively and inconsistently organized, and highly labor-intensive in authoring. This may explain the humorously-labeled Lost In Space problem of many of the first-generation hypertext information systems (Bevilaqua 1989 <refs -bevil>; Halasz 1988 <refs -halasz>). Without clear and consistent structure or visible "navigation" bearings, users easily lose position orientation or "direction" in the information base (Conklin 1987, p.40 <refs 6 1>). Attempts to author complex hypertext systems authoring using simple, intuitive, "seat of the pants" linking methods are likely to produce "spaghetti hypertext." Such systems are tangles of information threads that will confuse and disorient users (Urr 1991) <refs-urr>. This kind of approach would be analogous to trying to produce a substantial library catalog or indexing operation without using consistent procedures, vocabulary control, or cross references. Hypertext pioneer van Dam (1988) <refs 24 5> echoed these concerns, agreeing that being "lost in space" was a major problem. As he phrased it: My point of view is that in a sense hypertext gives us a goto, and a goto, as we all know, produces spaghetti. . . . [W]e need new forms, new flow of control kinds of constructs besides just unbridled gotoness of directed graph structures. . . . The other navigation aid we need is a tell-me-what-you've-got, . . . to derive new knowledge from old. We need people who. . .can think about classification and indexing and how we put things together. (894-895) Hypertext technology can provide simple, direct, efficient manipulation of text-based information. Hypertext can thus be used to essentially emulate, automate, and optimize the traditional print-based information access models, especially when combined with effective visual or graphic presentations. The efficiency and "leverage" of electronic format manipulation also multiplies both author and user efficiency in locating, identifying, incorporating, and utilizing information. Information system designers may thus profitably develop hypertext information access methods by using simulations of established print-based presentations. These may include displays of indexes, tables of contents, glossaries, or abstract summaries. These "print emulations" will incorporate hypertext associative links. Users will be able to use the embedded links for instant retrieval and display of referenced material. They can also use links for automatic execution of automated procedures. The hypertext linking can thus transparently interface to alternate information retrieval mechanisms, such as external bibliographic or full-text search software. This study will identify and recommend traditional information access models that can be profitably used in hypertext system authoring. Definitions Associative retrieval - "selection by association," the information access concept introduced by Vannevar Bush (1945) <refs -bush>, as an alternative to traditional classification and indexing access models. Hypertext - a computer information system containing nodes or units of text, graphics, sound, animation, video or procedural sequences, maintained in a linked network structure. Visible hypertext links or "buttons" are "associated" with the related information items. Link activation results in immediate display of the associated material, or execution of referenced functions. Hypertext links - highlighted or otherwise visible screen display markers or pointers, used to indicate or "point to" linked system nodes. Information access model - the conceptual model of the general access systems and control mechanisms used in traditional information access methodologies. Information delivery - mechanism for user presentation of a referenced source text or graphic image. Information filtering - traditional information access approaches such as bibliographic records, indexes, or classifications enable the user to "narrow down" or identify material of probable interest within mass archival collections. These traditional methods have generally used a prepared document surrogate and query mechanism, in order to access graphic documents containing desired information. Print-based methodologies - traditional print or computerized information retrieval systems, based upon formatted or manipulated displays of text. These include catalogs, bibliographic classifications, controlled vocabularies, indexes, abstracts, and bibliographic or full-text databases. These are all potential candidates for functional emulation using hypertext. Research Questions Pertaining specifically to the specific case study subject hypertext system: 1) What traditional information organization and access or retrieval methodologies currently exist in the subject hypertext information delivery application? 2) Which of the additional methodologies could effectively be emulated or achieved within the subject hypertext system? 3) What additional traditional, print-based information organization and access methodologies might be productive for incorporation into the generic hypertext medium? 4) What existing types of computer-based information retrieval systems might be useful for transfer or incorporation into hypertext authoring systems? (e.g., computerized indexing systems, thesaurus software, or full- text retrieval software.) Limitations 1) This is an exploratory case study, an initial effort at describing the transferability of proven information access and control methods to the new hypertext information medium. No studies have yet directly addressed this area. 2) The study will be limited to a single commercial hypertext authoring system implementation product. 3) The subject hypertext system operation is limited to the widely distributed MS-DOS microcomputer operating environment.