HCI Bibliography Home | HCI Conferences | ECHT Archive | Detailed Records | RefWorks | EndNote | Hide Abstracts
ECHT Tables of Contents: 90929494X

Proceedings of the ECHT'90 European Conference on Hypertext

Fullname:ECHT'90: European Conference on Hypertext
Note:Hypertext: Concepts, Systems and Applications
Editors:Antoine Rizk; Norbert Streitz; Jacques Andre
Location:Versailles, France
Dates:1990-Nov-27 to 1990-Nov-30
Publisher:Cambridge University Press
Standard No:ISBN 0-521-40517-3; hcibib: ECHT90
  1. Keynote Address
  2. Toolkits for Hypermedia Applications
  3. Formal Models and Query Languages
  4. Databases, Indices and Normative Knowledge
  5. Argumentation, Design and Knowledge Acquisition
  6. Turning Text into Hypertext
  7. Designing and Reading Hyperdocuments
  8. Navigation and Browsing
  9. Building Hypertext Applications
  10. Panels

Keynote Address

Assessing the Quality of Hypertext Documents BIBAK 1-12
  P. J. Brown
The greatest need in hypertext today is not for further technical wizardry, but for authors who can exploit the medium successfully.
   The paper discusses how we can help novices to become effective hypertext authors. Being effective not only means producing a hyperdocument of immediate appeal to readers; it also means ensuring that hyperdocuments have a potentially long lifetime, and can be maintained and enhanced by other authors over the years. Furthermore it means ensuring that the structure of a hyperdocument is 'correct' according to certain rules.
   In discussing these issues, the paper tries to draw lessons not only for authors but also for hypertext developers and researchers.
Keywords: Hypertext authorship, Hypertext refereeing, Virus, Testing, Integration, Maintenance

Toolkits for Hypermedia Applications

Building Hypertext on a Multimedia Toolkit: An Overview of Andrew Toolkit Hypermedia Facilities BIBAK 13-24
  Mark Sherman; Wilfred J. Hansen; Michael McInerny; Tom Neuendorffer
This paper discusses several hypermedia facilities built on top of the Andrew Toolkit (ATK) and their use in ATK applications. As a general-purpose, multimedia, application-development system, ATK permits many kinds of links, references and other connections to be made within pieces of content and between pieces of content, regardless of the content's medium. We argue that starting with a multimedia architecture facilitates the construction of all forms of hypermedia. Four specific hypermedia facilities implemented with ATK are discussed: an integrated web and indexing system (Help), a simple multimedia link facility (Link), a cross reference (Textref) capability, and a link-supporting embedded object language (Ness). As a toolkit, ATK is used to build other applications which inherit ATK's hypermedia facilities. Therefore we consider briefly the way that hypermedia facilities are used in conventional applications, such as mail systems.
Keywords: Hypertext, Implementation, Multimedia
The Toolkit Approach to Hypermedia BIBAK 25-37
  J. J. Puttress; N. M. Guimaraes
Hypermedia systems are usually developed as a single, self-contained application making the system specialized and difficult to retool for other applications. Thus each system requires the redevelopment of hypermedia storage and display capabilities. As long as developers need to write underlying hypermedia mechanisms, it remains impractical to extend hypermedia to many new domains.
   Our effort focuses on developing a toolkit that can be used by developers to add hypermedia functionality to their tools, independent of their particular application or environment. By doing this we gain flexibility and provide a common framework allowing applications to share information and user interface capabilities.
   The toolkit provides a simple hypermedia data model and an object-oriented user interface. The data model stores the underlying hypermedia structure and the application's data. The user interface model separates the data and the view components of the objects, which permits the construction of interfaces independent of the final display platform. The combination of these two components provides a powerful toolkit capable of either adding a touch of hypermedia to an application or constructing a monolithic hypermedia system.
   We describe the constraints of our development environments, our toolkit, and some typical applications of the toolkit, as well as our future plans.
Keywords: Hypermedia, User interface, CASE
Scenario-Based Hypermedia: A Model and a System BIBAK 38-51
  Ryuichi Ogawa; Hiroaki Harada; Asao Kaneko
This paper discusses an extended hypermedia model which presents data according to timed scripts. The model is intended to expand the notion of nodes and links, so that hypermedia systems can easily handle time-based, media-composite data objects, including audio and video. It provides script-based nodes to present text-graphic and audio-visual data synchronously, and buttons to control the data sequence. Based on this model, a prototype hypermedia system Videobook has been implemented and used experimentally. This paper also discusses the authoring issues and educational applications developed on Videobook.
Keywords: Multimedia, Scene node, Trigger button, Script, Videobook

Formal Models and Query Languages

A Hypertext Model Supporting Query Mechanisms BIBAK 52-66
  Foto Afrati; Constantinos D. Koutras
A formal model of hypertext is described in this paper. The purpose is two-fold: a) building on Garg's work ([Garg88]), to enhance his model so as to express advanced features of hypertext systems (such as structured nodes, scripts, typed and attributed links), and b) to demonstrate the expressive power of the model by showing that several problems concerning knowledge organization, browsing and navigation in the hyperspace, may have a simple solution in the framework of the model, as a result of a powerful query mechanism. The principles and fundamentals of the model are defined in detail, while its power and simplicity is illustrated by presenting some simple examples of information organization in hypertext fashion.
Keywords: Hypertext, Node, Link, Script, Button, Region
A Logical Query Language for Hypertext Systems BIBAK 67-80
  Catriel Beeri; Yoram Kornatzky
The search capabilities of hypertext systems are currently limited to retrieving collections of nodes and links based on predicates on their attributes and contents. To support sophisticated applications and user-tailored views of a hypertext document, we need a query language able to retrieve parts of a hypertext based on a specification of their structure. We present a logical query language permitting the formulation of such structural queries over hypertext. While the language is propositional, it includes a general notion of quantifier of the form appropriate for hypertext networks. Quantifiers are used for expressing formulas of the form: "For most paths from the current node, claim X holds". In particular, most quantified assertions in natural language are directly represented in our logic. Formulas in the language are used for a declarative definition of sophisticated user-tailored views of a hypertext document.
Keywords: Query languages, Views, Generalized quantifiers
A Model for Hypertext-Based Information Retrieval BIBAK 81-94
  Dario Lucarella
This paper approaches the problem of information retrieval from hypertext. In this context, the retrieval process is regarded as a process of inference that can be carried out either by the user exploring the hypertext network (browsing), or by the system, exploiting the hypertext network as a knowledge base (searching). That is the reason why a comprehensive model should take into account both of the perspectives, combining effectively browsing and searching in a unified framework. In the following, such a model is defined and implementation issues are outlined for a hypertext-based information retrieval system.
Keywords: Hypertext models, Information retrieval, Knowledge-based systems, Intelligent searching, Plausible reasoning

Databases, Indices and Normative Knowledge

HyperBase: A Hypermedia Engine Based on a Relational Database Management System BIBAK 95-108
  Helge A. Schutt; Norbert A. Streitz
Hypertext systems are valuable tools for creating, (re-) structuring, and presenting information bases. Until now, little has been done with respect to the underlying data model and even less with respect to system support for such a model. This leads to a significant mismatch between sophisticated organizational structures at the user interface level and the actual storage of persistent objects in simple file systems. Therefore, we have developed a general data model for hypertext data and implemented that model with the help of a database system. Here we exploit the fairly complex functionality of a commercially available relational database management system to implement a general purpose hypermedia engine which we call HyperBase.
Keywords: Datamodels of hypertext, Hypertext abstract machines, DBMS support for hypertext systems
Hyperindices: A Novel Aid for Searching in Hypermedia BIBAK 109-122
  Peter D. Bruza
In this article the formal basis of hyperindices is given. Hyperindices are a new means for supporting effective search in hypermedia. The basis of the hyperindex, the so called index expression, is treated in detail. It is shown how the hyperindex can be constructed using the structural properties of the index expression. The hyperindex is placed in a general framework for indexes which features quantitative and qualitative criteria with which index effectiveness can be judged.
Keywords: Hypertext, Information retrieval, Indexing
Intelligent Hypertext for Normative Knowledge in Engineering BIBAK 123-136
  Daniel Schwabe; Bruno Feijo; Werther G. Krause
We present a system that combines hypertext with a semantic representation of engineering norms. Since the representation is done via a Prolog encoding of an And/Or Graph, it is possible to discuss the relation between the execution of the (representation of the) norm and navigation in the hypertext. The system incorporates an interpretation of the norm by experts, and it is shown how this interpretation can be regarded also as an hyperview onto the hypertext.
Keywords: Norms, Prolog, Knowledge based systems

Argumentation, Design and Knowledge Acquisition

Author's Argumentation Assistant (AAA): A Hypertext-Based Authoring Tool for Argumentative Texts BIBAK 137-151
  Wolfgang Schuler; John B. Smith
We present the conceptualization and implementation of AAA, a prototype authoring tool for creating argumentation-based hyperdocuments. AAA is part of a more comprehensive effort of GMD-IPSI, where the hypertext authoring system SEPIA (Structured Elicitation and Processing of Ideas for Authoring) is developed. AAA will be used for writing and design experiments the results of which will be used in the design of SEPIA. It is designed to support the creation of argumentation patterns in accordance with the IBIS/PHI (Procedural Hierarchical IBIS) model combined with a micro argumentation structure according to Toulmin. For rapid prototyping purposes it has been implemented as a hypertext system using the Writing Environment WE developed at UNC.
   AAA uses a combination of different independent but cooperating modes of operation dedicated to different cognitive tasks of the argumentative writing process. The entire argumentation structure is represented as a layered network of typed nodes and links in which different layers correspond to different levels of abstraction.
Keywords: Authoring system, Hypertext application, Argumentation model
PHIDIAS: Integrating CAD Graphics into Dynamic Hypertext BIBAK 152-165
  Raymond J. McCall; Patrick R. Bennett; Peter S. D'Oronzio; Jonathan L. Ostwald; Frank M., III Shipman; Nathan F. Wallace
PHIDIAS is a hypermedia system for supporting environmental design. It embodies a theory of design as continual alternation between two complementary activities: construction of solution form and argumentation about construction. To support these activities it implements a number of advanced hypermedia concepts. These include an applicative query language providing search by both structure and content, virtual structures, composite graphic nodes, query-based graphic clustering, and "triggered" queries which connect construction acts to relevant sections of the argumentative network. PHIDIAS constitutes a new type of integrated information environment for design.
Keywords: Hypermedia, CAD, IBIS, Virtual structures
An Integrated Approach of Knowledge Acquisition by the Hypertext System CONCORDE BIBAK 166-179
  Martin Hofmann; Uwe Schreiweis; Horst Langendorfer
Knowledge acquisition is a crucial issue in developing expert systems. We describe a particular prototype of a hypertext system called CONCORDE that is able to support the entire process of knowledge acquisition. Since hypertext can be seen as a special structuring of various information units plus a fitting form of presentation, the data model of CONCORDE is discussed as well as its graphical browser.
Keywords: Knowledge acquisition, Typed links, CONCORDE

Turning Text into Hypertext

Hierarchy, Composition, Scripting Languages, and Translators for Structured Hypertext BIBAK 180-193
  P. David Stotts; Richard Furuta
In this paper we describe a hypertext translator-generator system that uses XTed, the visual Petri net editor from the {Chi}Trellis hypertext system, to specify the semantic component of a string-to-graph translation. XTed-specified parsers convert general authoring notations into structured {Chi}Trellis documents for browsing. The operative mechanism is termed a pair grammar, in which a string grammar and a graph grammar are paired in a one-to-one correspondence. When an XTed-specified parser reduces by one of its string grammar productions, the corresponding production in its graph grammar is used to generate a portion of the Petri net that implements that syntax. The use of pair grammars in conjunction with the Trellis model results in a general method of defining hypertext structure that is both hierarchical and compositional.
Keywords: Hierarchy, Graph grammar, Translation, Trellis, Petri nets
Links and Structures in Hypertext Databases for Law BIBAK 194-211
  Eve Wilson
This paper considers the need to automate the conversion of traditional legal documents into an integrated hypertext database. It describes how this has been done at the University of Kent at Canterbury by the Justus suite of programs, which converts legal documents of diverse structures to run under the Guide hypertext system. It discusses the types of links the system uses to cater for: 1. linear and hierarchical structures; 2. directed graphs; 3. annotational or associative links; 4. index or concept links.
   It then discusses how these links can create different virtual structures for the document collection to give flexibility of access and navigation. Some of these mirror structures which exist in traditional paper systems; others are unique to computerised systems. Next it considers the use of embedded expert systems to steer variable paths through the documents. Finally it summarises the achievements so far, and the goals ahead.
Keywords: Hypertext links, Hypertext structures, Information retrieval, Law
An Apprentice that Discovers Hypertext Links BIBAK 212-223
  Mark Bernstein
A simple apprentice that proposes new content links in hypertexts has been added to the Hypergate hypertext environment, and its behaviour and performance are evaluated under realistic conditions. Despite a fundamentally lexical approach, the link apprentice finds a significant number of intriguing and meaningful links very quickly and without substantial overhead.
Keywords: Indexing, Authoring systems, Apprentices, Tours, Information retrieval

Designing and Reading Hyperdocuments

Towards an Aesthetics of Hypertext Systems. A Semiotic Approach BIBAK 224-237
  Peter B. Andersen
In hypertext the reader can choose between several reading paths, and this is an essential characteristic of the medium. Composing paths that give meaningful readings is an important task for the hypertext author, but so far very little is known of the "rhetorics of paths". Based on the practical work of evaluating hypertext exercises written by students, the paper gives examples of errors of composition and sets up a conceptual framework, borrowed from semiotics, for discussing aesthetic issues of hypertext design. The concepts are intended to make it easier for the author to design the product from a reader's point of view, and an authoring tool supporting these concepts is sketched. Finally, the paper illustrates how insights from text stylistics and film aesthetics can be usefully exploited in the field of hypertext.
Keywords: Models of hypertext, Authoring systems, Models of reading and writing, Aesthetics, Semiotics
Linking Considered Harmful BIBAK 238-249
  Laura De Young
Arbitrary linking of data in hypertext allows for great flexibility, but the result is often hypertext in which users readily become disoriented. Where possible, it is desirable to provide support for structuring hypertext in a way that makes it easier to organize and understand. This can be done by identifying the underlying structure of the ways specific sets of data are related. Providing support for use of such structures in a hypertext system may yield benefits similar to those found in using higher-level programming constructs in programs.
Keywords: Hypermedia, Structure, Disorientation
Interactive Text Processing by Inexperienced (Hyper-) Readers BIBAK 250-360
  Jean-Francois Rouet
This paper focuses on the development of cognitive strategies in secondary school students, when learning to use electronic nonlinear documents. We study the costs and benefits of learning through nonlinear texts, from a psycholinguistic point of view. In the course of two experimental sessions, 148 11-12 and 13-14 year-old secondary school students were trained to use an interactive text-presentation software. Parameters such as the expression of relationships, selection marking, and pagination influenced both local and global aspects of nonlinear reading. Training improved orientation strategies at both academic levels. Implications of these results for the design of instructional nonlinear documents are discussed.
Keywords: Cognitive strategy, Instructional text, Interface organization, Orientation, Training

Navigation and Browsing

Non-Intrusive Hypertext Anchors and Individual Colour Markings BIBAK 261-273
  W. J. Irler; G. Barbieri
One interface feature of the hypertext systems is the visualization of the link anchors, often in form of buttons which invite to click on. This functional information increases the cognitive load while reading a text on the computer. Comprehension of the presented material competes with the orientation in the hypertext. The paper treats the cognitive arguments, and explores interface and back-end questions related to this problem. A browsing model is developed which proposes a button-less screen, a click-anywhere suggestion, and pop-up local maps as a selection and preview mechanism. Links are finally activated by mouse movements towards a destination area, rather than by still button pressing. As a way to individualize and re-structure the visual aspect of the pages, the reader can permanently mark selected words with colour. Marked words may be added to a keyword index which is linked back to the text. An experimental implementation uses ToolBook, a scripted construction set in MS-Windows 3.0. The design issues and some results are described for an educational application concerning pendulums.
Keywords: Invisible anchors, Text marking, Browsing model
SaTellite: Hypermedia Navigation by Affinity BIBA 274-287
  Xavier Pintado; Dennis Tsichritzis
SaTellite is an exploration tool for a hypermedia environment. Navigation support is based on the concept of affinity between objects; that is, a relationship with an associated intensity. The user is presented with a two dimensional representation that provides a view of the hypermedia environment where objects lying close together have some affinity in a sense defined by a chosen measure. The system provides for multiple views by modifying the underlying measure of affinity. SaTellite promotes hypermedia navigation by context-dependent proximity.
   Based on the affinity concept, we develop a dynamic layout algorithm for the display of views. Because the views are based on affinity information, the links of the hypergraph do not appear explicitly.
Browsing in Hyperdocuments with the Assistance of a Neural Network BIBAK 288-297
  Frederique Biennier; Michel Guivarch; Jean-Marie Pinon
The high degree of freedom a user has to browse through an hyperdocument often makes him puzzled. His main problems are first the expression of his informal need, sometimes using ideas associations, then finding the path in the hyperspace to reach relevant information.
   The guiding system proposed in this paper enriches the hyperdocument structure with undirect semantic links, i.e., associations between nodes according to their contents. Nodes' contents are connected to multimedia keywords called tags; direct and reverse associations between nodes and tags are embedded in a bidirectional neural network which allows inductive retrieval. One user controls the process thanks to some simple parameters: specialization level for selected nodes, precision of tags, inertia selector, tolerance functions for specialization and precision spreads.
   Upon request, the system dynamically raises a path that organizes the results of a query, contextual or not, adaptatively taking into account users' profile and special needs. Weights in the neural network may also be slightly corrected from experience, adapting the association capability to users on their average.
Keywords: Semantic browsing, Neural network, Dynamic path

Building Hypertext Applications

MICROCOSM: An Open Model for Hypermedia with Dynamic Linking BIBAK 298-311
  Andrew M. Fountain; Wendy Hall; Ian Heath; Hugh C. Davis
There are currently a number of commercially available hypertext and hypermedia systems, of varying levels of sophistication and usability, but there are still many problems to be resolved in the design of such systems. In this paper, we itemise some of the major problems that we have identified as possibly causing a barrier to the growth and development of hypermedia applications outside the research community. A model of an open hypermedia architecture with dynamic linking features is proposed that moves some way to resolving these problems, and the first implementation of the system, Microcosm, is presented and discussed.
Keywords: Hypertext, Multimedia, Dynamic linking
Inside Macintosh as Hypertext BIBAK 312-323
  Brian L. Bechtel
SpInside Macintosh is a hypertext compact-disc version of Inside Macintosh, the multi-volume programmer's reference work for the Macintosh family of computers. SpInside Macintosh is implemented in HyperCard and distributed on CD-ROM. We describe the design decisions and principles we followed in creating SpInside Macintosh. We also give user feedback and a retrospective critique of the design.
Keywords: Inside Macintosh, CD-ROM, Hypertext
Hypertext from the Data Point of View: Paths and Links in the Perseus Project BIBAK 324-336
  Elli Mylonas; Sebastian Heath
The Perseus Project is building a system for studying Classical Greece, incorporating into it several different types of source material. In order to minimize system development time, and to make accessible to the users of the system faster, it is being developed on Macintosh computers, using HyperCard. This paper describes two navigational methods that have been created in Perseus: generalized linking, and paths. They were chosen because they could provide the most flexibility and the most functionality. Their implementation is briefly described, as well.
Keywords: Linking, Navigation, Path, Perseus project


How Should Hypermedia Authoring Systems for Computer Aided Instruction Look Like? BIB 337-342
  Peter A. Gloor; Michael R. Kibby; Ray McAleese; Max Mulhauser; Gerald C. Nelson; Daniel Russel
The Advantages of Hypertext for Large Information Spaces; Where Are the Big Systems? BIB 343-346
  Patricia Baird; Jacqueline Covo; Ben Shneiderman; Ian Williams; Renee Deter
Hypertext and Electronic Publishing BIBAK 347-353
  Richard Furuta; Heather Brown; Steven R. Newcomb; Roberto Minio; Vincent Quint; Roy Rada; Laurence A. Welsch
The underlying models of hypertext and of structured documents are quite similar in many ways. The charter of this panel is to examine the similarities and differences between the models and to consider the applicability of document standards such as SGML, ODA, etc., to hypertext.
Keywords: Hypertext models and standards, Structured document models and standards
What's Specific about User-Interfaces for Hypertext Systems? BIB 354-361
  Norbert A. Streitz; Janet Walker; John A. Waterworth; Patricia Wright; Randall H. Trigg
Hypertext and Information Retrieval: What are the Fundamental Concepts? BIBA 362-366
  W. Bruce Croft; Nicholas Belkin; Marie-France Bruandet; Rainer Kuhlen; Tim Oren
Both hypertext and information retrieval (IR) systems provide access to databases consisting primarily of text documents. Both types of systems structure the content of these documents and support interaction with the users in order to improve the effectiveness of retrieval. Despite these similarities, hypertext and IR are generally regarded as separate research areas, with some overlap, but essentially different research agendas. To clarify these differences as well as the areas of overlap, the members of this panel will attempt to define the fundamental concepts and the major research issues in each area, with special emphasis on their own research.
Strategic Issues in European Hypertext Research and Development BIB 367-369
  Flavio Argentesi; N. Streitz; R. Hansen; G. Degli Antoni; A. Cicu