Figure 1: NESTOR desktop showing various toolbar options
John Eklund, Access Australia Co-operative Multimedia Centre, Technology Park, Cornwallis Street Eveleigh, j.eklund@cmcaccess.com.au
James Sawers, The Institute of
Interactive Multimedia, The
University Of Technology, Sydney,
james.sawers@uts.edu.au
Romain Zeiliger,
National Scientific Research Centre
(CNRS), GATE, 93 ch. des Mouilles, 69130 Ecully, FRANCE,
zeiliger@irpeacs.fr
This paper describes NESTOR NAVIGATOR, a graphical Web browser
which supports collaborative learning (Zeiliger, Belisle &
Cerratto, 1999 in press). NESTOR is based on the principle that an
individual's path through an information space reflects their
discourse context with the information, it allows them to personalise
that space, and in effect, solve their own navigation problems. In a
window to one side of the browser window, it dynamically builds a
navigable overview map of the hyperspace as the user interacts with
it. It distinguishes the type of node that the user is currently at
and illustrates possible paths from that node. Users are able to
annotate nodes with personal notes, and describe their preferred path
through the information space, in other words, which path contains
the most meaning for them. Users may export these descriptions as
'tours' and thus share their personal interactions with the
information space with others. In terms of solving navigation
problems, NESTOR's approach is to provide an interactive, stimulating
environment where the learner's expertise is deployed, rather than
drawing on knowledge held in some expert model as in a
knowledge-based system. It encourages users to reflect on their
interactions with an information space, to augment those interactions
with annotations, to collaborate with others through the sharing of
tours and annotated maps, and to apply their own methodologies to
solve navigational problems. We overview the software, with
particular emphasis on the principles of discourse theory on which it
is based, and outline ways in which it is being integrated into
teaching and learning contexts, and describe experiments that are
being undertaken with it.
Most navigation on the Web follows one of the two primary
interaction modes of browsing and simple query. In the context of an
educational use of hypermedia rather than for simple information
retrieval, the relationship between the content of the hypermedia
nodes and navigation through those nodes is critically important as
it reflects an individual content structure. Yet the common browser
makes no effort to make a user's 'transaction' with the space more
relevant by recognising the content of component documents.
Höök & Svensson (1998) suggest that the metaphor
"navigation in an information space" involves traditional way-finding
activities to a known destination, as well as exploration and object
identification (Benyon & Höök, 1997). Way-finding
implies that users of an educational hypermedia system have some
learning goals, and their task is to move through the information
space in a manner which is the most efficient, both in a navigational
sense and in terms of attaining their learning objectives. Linard
& Zeiliger (1995) make the point that navigation occurs at two
levels: both in the interface and in the content. There is an
important distinction between these two types of navigation which is
often not made clear when discussing navigation issues in hypermedia.
Navigation in an interface involves users becoming oriented to the
navigational tool, such as the basic functioning of a Web browser,
while navigating a knowledge space involves users having some
understanding of the topography of the domain in which they are
immersed, what content is in certain sections, how it relates to
other sections and how it may be traversed. Other writers believe
that a hypertext system can be examined on three levels
(Germán & Cowan, 1997 [HREF1]),
namely the way it is presented, the way it can be navigated, and
finally its semantic structure. Consider a semantic network where
nodes are knowledge units linked to others in a logical fashion. If
the user has some understanding of the content of the current node
and its relation to surrounding nodes, this situates them in the
hyperspace (Zhao, O'Shea & Fung, 1993). Navigation is clearly
movement which involves a decision-making process, and as such is
inseparable from a consideration of context (Jul & Furnas, 1997).
Jul & Furnas (1997) identify three related levels of structure:
the inherent structure of the information, a structure imposed by the
author, and finally the users' view of the material which reflects
the structure of the knowledge in their mind.
Constructivism is a commonly used term to mean that learning occurs
through an individual construction of meaning and understandings. In
the cognitive domain, the idea that learning is a sequence of
transactions between object and subject is found in Discourse Theory.
This research area is directly relevant to navigation in hypertext
systems (Inder & Oberlander, 1994 [HREF2]).
As Eklund & Woo (1998) note, theories of discourse structure are
designed to model the construction and evolution of the structures
underlying extended dialogues. Drawn from the field of natural
language processing, the focus has been on the creation and analysis
of coherent discourse (Grosz & Sidner, 1976; Hobbs, 1979; Mann
& Thompson, 1987; Reichman, 1979). Theories have been offered
which attribute a hierarchical structure to discourse, and relations
of various types between the discourse segments. The theories
attribute certain characteristics to links between the segments: both
in terms of the user's "reasons" for making the connections and how
these connections make some personal sense to the user. These
theories help to determine which structures and knowledge sources are
relevant in either interpreting or generating a connected discourse
which is meaningful for an individual - a users' discourse with the
knowledge is defined by the particular path they take through it, and
this is facilitated by navigation tools which acknowledge and
annotate the individuality of those paths (see Nielsen, 1999).
For example, static navigation features such as "next" or "continue"
provide one path through the hyperspace and thus present a static
discourse.
As learners may construct a great number of varying paths through a
hypertext system, they have different interactions or discourses with
the knowledge. The choice of navigation may be influenced by both
task and document structure (Wright & Lickorish, 1990), as well
as the prior knowledge and intent of the learner. This variable may
be controlled in a closed corpus context, but the Web at large has
been described as an 'unmoderated space' (Jul & Furnas, 1997) as
no formal domain structure has been imposed on the component parts of
a users' transaction with a typical subspace. But it remains that
reading Web documents and thus navigating them relies - more than in
the case of other media - on a personal construction of what is
meaningful. Accordingly, we propose that a software tool with the aim
to support navigation should be first a constructivist environment
i.e. an environment which provides means for gathering, representing
or externalizing (Vygotsky, 1962), structuring and creating
navigational objects.
This point is even more valid in the educational domain as
learners are accustomed to accessing information spaces such as
libraries which are not only strongly moderated but whose categories
parallel their formal educational experiences. It is likely that when
learners explore the Web they have to face a new problem: it conveys
categories which conflict with that from their formal education, as
it provides less structure and more navigational possibilities. Our
approach is that such a structure must be imposed by the
individual and at the time they encounter it,
after which it may be shared with others.
NESTOR Navigator [HREF3]
is a Web browser which is built on the principle that an individual's
path through an information space reflects their discourse context
with the information, it allows them to personalise that space, and
in effect, solve their own navigation problems. In a window to one
side of the browser window, it dynamically builds a navigable
overview map of the hyperspace as the user interacts with it. It
distinguishes the type of node that the user is currently at and
illustrates possible paths from that node. Users are able to annotate
nodes with personal notes, and describe their preferred path through
the information space, in other words, which path contains the most
meaning for them. Users may export these descriptions as 'tours' and
thus share their personal interactions with the information space
with others. In terms of solving navigation problems, NESTOR's
approach is to provide an interactive, stimulating environment where
the learner's expertise is deployed, rather than drawing on knowledge
held in some expert model as in a system offering adaptive navigation
support (Brusilovsky, Eklund & Schwarz, 1997). It encourages
users to reflect on their interactions with an information space, to
augment those interactions with annotations, to collaborate with
others through the sharing of tours and annotated maps, and to apply
their own methodologies to solve navigational problems.
NESTOR Navigator (currently at version 5.3) runs under the Windows95 or NT 4 operating systems and requires Internet Explorer 4.01 or later. It provides basic browsing capabilities, a personal home page customisable by each user, annotation of hyperlinks, collection and storage of notes from hypertext pages and a drag and drop interface for the structuring and making sense of personal and public hypertext spaces. NESTOR allows for the saving and collaborative sharing of these mapped spaces in the NESTOR 'maps' and 'bags', which are structured containers for hyperlinks, concepts, annotations and related documents.
The NESTOR desktop shows (Figure 1) some of the various toolbar options and the users 'Project/home page'. Each configured user (in this case James) has a personal homepage defined which is fully editable using the HTML capabilities of IE 4 and can become the basis for each user's personal Web.
Figure 1: NESTOR desktop showing various toolbar options
Figure 2: Configuring NESTOR for Collaborative work
As a hypertext space is traversed in the integral browser, NESTOR creates a visual map of the 'journey' in the left-hand pane. This constructed map can include hyperlinks, annotations, keywords and personal documents. Several hyperlinks have been added to James' map in this case, the last node of which is visible in the right-hand browser pane. These hypertext spaces can be annotated and saved as '.map' files for sharing with others. Figure 3 shows the beginning the mapping of a hypertext space.
Figure 3: Beginning the mapping of a hypertext space
Figure 4: Mapping an additional hyperspace and adding a NESTOR keyword 'concept' to a hypertext node.
Figure 5: Adding a further mapping and a personal annotation.
NESTOR can also be used to produce maps of complex domains with included concept keywords and annotations. These maps can be shared with others as guided tours of key concepts and materials related to these domains.
Figure 6. Advanced mapping of a cognitive domain. The guided tour control panel allows traversal of saved hypertext spaces by others.
As a further collaboration feature NESTOR allows for the collection and storage of chunks of information gathered from HTML pages in a repository called the 'Bag'. The content is formatted as HTML and can be saved to a file for inclusion in other documents or for sharing with others.
Figure 7. Publishing information from the NESTOR 'bag' to the preconfigured ftp site for download by others.
The evaluation process focuses on one hand on constructive navigation as a new approach to Web navigation support and on the other hand on NESTOR as a software product. It is currently taking place in the framework of several National and European projects and it is conducted in real educational settings. NESTOR is available for free on the Web and NESTOR users participate actively to its improvement and development. Here is a list of running projects which provide frameworks for NESTOR experiments. More detail of these projects can be found on the Web:
ORION, program ARASSH, supported by Région Rhone-Alpes, France,"Conception d'aides à la navigation dans les hypermedia sur les réseaux globaux"
Setting: 2 secondary schools in Lyon and StEtienne.
Users: schoolboys aged 15-16, 2 groups of 12, 3 sessions of 4 hours
Domain: Economics
http://herakles.imag.fr/Geod/pages_html/projets/ORION.html
LEARN-NETT, program SOCRATES, supported by European Commission DG22,
"Distance and Open Education: a Learning network for teachers."
Setting: 7 Universities in UK, Belgium, Switzerland, Spain.
Users: students, groups of 4 collaborating at a distance, 2 month session.
Domain: Educational technology
http://tecfa.unige.ch/proj/learnett/
The Net @ Company, in collaboration with School of Management EM-Lyon, France
Setting: Engineering school.
Users: students, 3 groups of 24, 3 months sessions.
Domain: Management
Participating students are knowledgeable in the use of computers and already familiar with Web navigation. Their task consists of finding information on the Web on a given topic. Then they have to structure the information they have gathered while at the same time they mix it with the documents they create themselves. This is mainly individual work but group collaboration would be necessary to produce a common final document which takes the form of a NESTOR map.
First results show a decrease in user's "surface" disorientation: when compared with published data on navigation, statistics computed from the NESTOR traces show that users do not go 'home' as often, do not rely heavily on the "back" command and tend to make longer traversals. However, traces as well as direct observations reveal a serious disorientation in Web contents: young navigators in school don't "see" the answer to their questions while it is displayed right on their screen in the current document; this might be explained in terms of categorisation problems and lack of "off-screen" context which prevents a deep interpretation of the text displayed.
NESTOR is a Web client which runs on Microsoft Windows 95/98/NT platforms. It presents as a stand-alone Web browser, however the basic browsing and HTML editing features are enabled by two of Microsoft's Active-X components: TWebBrowser for browsing and THTMLEd for editing. That means that NESTOR takes advantage of InternetExplorer 4.01 retrieval motor, options (such as cache or connection options for example), popup menus and so on. NESTOR makes also extensive use of Dynamic HTML. The unique NESTOR left-window interface deals with the construction of maps and of a personal information space. In that view, NESTOR implementation can be considered as constituting an add-on to standard Internet browsers.
The main screen of NESTOR is divided into two windows: a standard browsing window (to the right) and a map window (to the left). While both windows support a range of navigational activities, the map window has been designed to facilitate information-structuring activities. Direct manipulation prevails in the map window while hypertext navigation is the standard mode in the browsing-window. The map window can be thought of - at first - as a dynamic map representing the visited Web subspace. Most navigation operations (such as backtracking) are available in both windows and provide a crossed feedback. The map provides a direct access to every visited document. NESTOR also records the raw navigation history and allows for re-constructing a map from a selection of this history. As "re-visit" generally accounts for more than 50% of users' operations, the map and history features (and their combinations) are very useful. NESTOR implements a special mapping for search motors (representing queries in order to facilitate the exploration of the filtered documents) and some handy features which facilitate directed searching like a "breadth-first" mechanism (Newfield et al, 1998).
NESTOR's approach is to provide an interactive, stimulating environment where the learner's expertise is deployed, rather than drawing on knowledge held in some expert model. It encourages users to reflect on their interactions with an information space, to augment those interactions with annotations, to collaborate with others through the sharing of tours and annotated maps, and to apply their own methodologies to solve navigational problems. In this paper we have highlighted the importance of a user's individual discourse with an information space reflecting a model of their underastanding of the information it contains, and have outlined discourse theory as a possible theoretical model of human learning which is supported by NESTOR's approach.
NESTOR [HREF3] features are best experienced in a demonstration. These features may be classified into seven bundles: a) representing self navigational experience, b) note-taking, c) constructing a personal Web, d) creating keywords objects and conceptual areas, e) creating and saving maps, f) using education oriented features such as "the bag" and the "guided tours", g) sharing bookmarks, bags, maps and tours with other users.
Benyon D & Hook K (1997) Navigation in Information Spaces: supporting the individual, In S. Howard, J. Hammond & G. Linedgaard (eds.). Human-Computer Interaction, INTERACT97. Chapman and Hall. p. 39-46.
Brusilovsky P, Eklund J & Schwarz E (1997) Adaptive Navigation Support in Educational Hypermedia on the World Wide Web. In Howard S, Hammond J & Linddgaard G (eds.). Human-Computer Interaction INTERACT97, The 6th IFIP World Conference on Human Computer Interaction. Sydney. Chapman & Hall, New York. p. 278-285.
Eklund J & Woo R (1998) A Cognitive Perspective for Designing Multimedia Learning Environments. In R Corderoy (ed.) Proceedings of ASCILITE98, Wollongong, December. The Printery, UOW. p. 181-190.
Esnault L & Zeiliger R (1999 in press) Learning, Teaching and Browsing with NESTOR. IRMA 1999 Conference, May 16-19, 1999, Hershey, Pe., USA.
Höök K & Svensson M (1998) Evaluating adaptive navigation support. In Nils Dahlback (ed.) Exploring Navigation; Towards a framework for design and evaluation of navigation in electronic spaces. SICS Technical Report T98:01. SICS. p. 141-151.
Jul & Furnas (1997) Navigation in Electronic Worlds, a CHI'97 workshop, SGCHI bulletin, Vol. 29, No. 4. ACM press.
Zeiliger R, Reggers T, Baldewyns L & Jans V (1997) Facilitating Web Navigation: Integrated Tools for Active and Co-operative Learners. In Halim Z, Ottmann T & Razak Z (eds.). Proceedings of ICCE97. AACE. p. 529-537.
Linard M & Zeiliger R (1995) Designing a navigational support for educational software. In B. Blumental & J. Gornostaev (eds.). Proceedings of EWHCI'95. Moscow: ICSTI. p. 130-145.
Nielsen J (1999) User Interface Directions for The Web. Communications of the ACM, Vol. 42, No. 1, p. 65-72. ACM Presss.
Newfield D, Sing Sethi B & Ryall K (1998) Scratchpad: Mechanisms for Better Navigation in Directed Web Searching. Proceedings of UIST Conference, ACM Press, p. 1-7.
Vygotsky L (1962) Thought and Language. MIT Press, Cambridge.
Zeiliger R (1998) Supporting Constructive Navigation of Web Space. In K Hook, A munro & D Benyon (eds.) Workshop on Personalised and Social Navigation in Information Space. SICS Technical Report T98:02. p. 91-101.
Zhao Z, O'Shea T & Fung P (1993) Visualization of semantic relations in hypertext systems. In T. Ottman & I.Tomek (eds.). Proceedings of ED-MEDIA'93, World conference on educational multimedia and hypermedia. Orlando, FL: AACE. p. 556-564.
Zeiliger R, Belisle C & Cerratto T (1999 in press) Implementing a Constructivist Approach to Web Navigation Support. Paper to be presented at ED-MEDIA'99, June 19-24, AACE, Seattle,Wa., USA.
John Eklund, James Sawers & Romain Zeiliger, © 1999. The authors assign to Southern Cross University and other educational and non-profit institutions a non-exclusive licence to use this document for personal use and in courses of instruction provided that the article is used in full and this copyright statement is reproduced. The author salso grant a non-exclusive licence to Southern Cross University to publish this document in full on the World Wide Web and on CD-ROM and in printed form with the conference papers and for the document to be published on mirrors on the World Wide Web.