Published on the Web from the WebNet-96
(forthcoming)
proceedings with permission of the Association for the Advancement of Computing in Education (AACE).
Peter Brusilovsky
School of Computer
Science
Carnegie Mellon University
Pittsburgh, PA 15213, USA
E-mail: plb@cs.cmu.eduElmar Schwarz and Gerhard
Weber
Department of Psychology,
University of
Trier
D-54286, Trier, Germany
E-mail: {schwarz | weber
}@cogpsy.uni-trier.de
Abstract:
An Electronic textbook is a popular kind of educational applications
on World
Wide Web (WWW). We claim that adaptivity is especially important for
WWW-based
educational applications which are expected to be used by very
different groups
of users without assistance of a human teacher . In this paper we
describe an
approach for developing adaptive electronic textbooks and present an
authoring
tool based on this approach which simplifies the development of
adaptive
electronic textbooks on WWW.
World Wide Web opens new ways of learning for many people. Now,
educational
programs and learning materials installed and supported in one place
can be
used by thousands of students from all over the world. However, most
of the
existing educational WWW applications use the simplest solutions and
are much
more limited than existing 'on-site' educational systems and tools.
For many
designers, an ideal format of educational WWW material seems to be a
static
electronic copy of a regular textbook: chapter by chapter, page by
page,
picture by picture. Such electronic textbooks are non-adaptive, i.e.,
students
with different abilities, knowledge, and background get the same
educational
material in the same form.
We claim that adaptivity is especially important for educational
programs on
WWW which are expected to be used by very different classes of users
without
assistance of a real teacher (who usually can provide adaptivity in a
normal
classroom). Currently, we can name very few adaptive educational
applications
on WWW [Brusilovsky, Schwarz & Weber 1996; Kay & Kummerfeld
1994; Lai,
Chen & Yuan 1995; Nakabayashi et al. 1995]. All these
applications keep a
model of the user between sessions and use it to adapt the teaching
sequence
and the presentation of the material to a given user. The problem is
that
adaptive electronic textbooks are "knowledge-rich" applications and
they are
not very easy to design. There are some authorinig tools for
developing
"static" electronic textbooks on WWW [Goldberg, Salari & Swoboda
1996;
Thimbleby 1996], but there are no tools available to support a
designer in
creating an adaptive textbook on WWW.
A possible approach for designing adaptive electronic textbooks on
WWW was
suggested recently in [Brusilovsky 1995]. This approach was further
elaborated
by the ELM research group in the University of Trier which applied it
for
developing an adaptive WWW-based LISP textbook ELM-ART [Brusilovsky,
Schwarz
& Weber 1996][Schwarz, Brusilovsky & Weber 1996]. Now we have
generalized the experience gained in ELM-ART
project and developed a subject-independent tool which simplifies the
process
of creating adaptive electronic textbooks on WWW. In this paper we
describe our
current approach for developing adaptive electronic textbooks on WWW
and
present an authoring tool which is based on this approach. In
addition, we
provide some recommendations for possible users of our tool and a
brief review
of relevant works.
Our approach to developing adaptive electronic textbooks on WWW based
on the
ideas from the areas of Intelligent Tutoring Systems [Wenger 1987]
and Adaptive
Hypermedia [Brusilovsky 1996]. Our adaptive textbooks use knowledge
about its
domain (represented in the form of domain model) and about its users
(represented in the form of individual user models). The domain model
serves as
a basis for structuring the content of an adaptive ET. We distinguish
two parts
in an adaptive ET: a glossary and a textbook. Both
these parts
are based on the domain model. The student model is used by an
adaptive ET to
adapt its behavior to each particular user.
According to our approach, the key to adaptivity in an adaptive ET
are the
domain model and the user model. The simplest form of domain model is
just a
set of domain concepts. By concepts we mean elementary pieces of
knowledge for
the given domain. Depending on the domain and the application area,
concepts
can represent bigger or smaller pieces of domain knowledge. A more
advanced
form of the domain model is a network with nodes corresponding to
domain
concepts and with links reflecting several kinds of relationships
between
concepts. This network represents the structure of the domain covered
by a
hypermedia system. The domain model provides a structure for
representation of
the user's knowledge of the subject. For each domain model concept,
an
individual user's knowledge model stores some value which is an
estimation of
the user knowledge level of this concept. This type of model (which
is called
an overlay model) is powerful and flexible: it can measure
independently
the user's knowledge of different topics.
The glossary is the central part of the ET. According to our
approach, the
glossary is considered as a visualized (and externalized) domain
network. Each
node of the domain network is represented by a node of the
hyperspace, while
the links between domain network nodes constitute main paths between
hyperspace
nodes. The structure of the glossary resembles the pedagogical
structure of the
domain knowledge and, vise versa, each glossary entry corresponds to
one of the
domain concepts. The links between domain model concepts constitute
navigation
paths between glossary entries. Thus, the structure of the manual
resembles the
pedagogic structure of the domain knowledge. In addition to providing
a
description of a concept, each glossary entry provides links to all
book
sections which introduce the concept. It means that the glossary
integrates
traditional features of an index and a glossary.
A human-written textbook represents human teaching expertise on how
to
introduce the domain concepts to the learners. It is usually a real
textbook
represented in hypermedia form. A textbook is hierarchically
structured into
units of different level: chapters, sections, and subsections. To
make the
textbook "more intelligent" and to connect it to the glossary, we
have to let
the system know what each unit of the textbook is about. It is done
by indexing
of textbook units with domain model concepts. For each unit, a list
of concepts
related with this unit is provided (we call this list spectrum
of the
unit). For each involved concept, the spectrum of the unit can
represents also
the role of the concept in the unit. Currently we support two roles:
each
concept can be either a outcome concept or a
prerequisite
concept. A concept is included in the spectrum as a outcome concept
if some
part of this page presents the piece of knowledge designated by the
concept. A
concept is included in the spectrum as a prerequisite concept if a
user has to
know this concept to understand the content of the page. Indexing is
a
relatively simple but powerful mechanism, because it provides the
system with
knowledge about the content of its pages: the system knows which
concepts are
presented on each page and which concepts have to be learned before
starting to
learn each page. It opens the way for several adaptation techniques
presented
in the next subsection.
The knowledge about the domain and about the textbook content is used
to serve
a well-structured hyperspace. The system supports sequential and
hierachical
links between section. It generates the table of content where all
entries are
clickable links. In addition, it generates links between the glossary
and the
textbook. Links are provided from each textbook unit to corresponding
glossary
pages for each involved concept. On the other hand, from each
glossary page
describing a concept the system provides links to all textbook units
which can
be used to learn this concept. These links are not stored in an
external format
but generated on-the-fly by a special module which takes into account
the
user's current state of knowledge represented by the user's model.
This
approach is not only reducing page design time but also provides room
for
adaptation. In particular, our approach supports two adaptation
techniques
which have been applied in ELM-ART [Brusilovsky, Schwarz & Weber
1996]:
adaptive navigation support and prerequisite-based help.
Our approach provides many more opportunities for browsing the course
materials
than traditional on-line textbooks. The negative side of it is that
there is a
higher risk for the user to get lost in this complex hyperspace. To
support the
user navigating through the course, the system uses adaptive
annotation,
an adaptive hypermedia [Brusilovsky 1996] technique. Adaptive
annotation means
that the system uses visual cues (icons, fonts, colors) to show the
type and
the educational state of each link. Using the user model, the system
can
distinguish several educational states for each page of material: the
content
of the page can be known to the user, ready to be learned, or not
ready to be
learned (the latter example means that some prerequisite knowledge is
not yet
learned). The icon and the font of each link presented to the user
are computed
dynamically from the individual user model. They always inform the
user about
the type and the educational state of the node behind the link. Red
means not
ready to be learned, green means ready and recommended, and white
means no new
information. A checkmark is added for already visited sections.
The system knowledge about the course material comprises knowledge
about what
the prerequisite concepts are for any unit of the textbook. Often,
when users
have problems with understanding some explanation or example or
solving a
problem, the reason is that some prerequisite material is not
understood well.
In that case they can request prerequisite-based help (using a
special
button) and, as an answer to help request, the system generate a list
of links
to all sections which present some information about background
concepts of the
current section. This list is adaptively sorted according to the
user's
knowledge represented in the user model: more "helpful" sections are
listed
first. Here "helpful" means how informative the section is to learn
about the
background concepts. For example, the section which presents
information about
an unknown background concept is more informative then a section
presenting
information about a known concept. The section which presents
information about
two unknown background concepts is more informative then a section
presenting
information about one concept.
Out tool (provisional name is InterBook) is aimed to help the author
to
transfer a normal textbook existing in electronic form into an
adaptive ET. To
get the most benefits from it an author should start creating an
electronic
textbook with a hierarchically structured MS Word file. This section
demonstrates a typical scenario of using this tool for the case when
the
original textbook is available as MS Word file.
Step 1. Creating the list of domain concepts. Before starting
to produce
an adaptive ET, an author has to think about the list of domain
concepts which
will be used to annotate pages. An author does not have to have the
list of
all concepts before starting the work, it could be made when
annotating
the book.
Step 2. Structuring and Annotation. To let InterBook recognize
the
structure of a book, an author has to use the regular way of
structuring an MS
Word file. It means that the titles of the highest level sections
should have a
pre-defined text style "Header 1", the titles of its subsections
should have a
pre-defined text style "Header 2", and so forth. Then an MS Word file
has to be
annotated and indexed by the course author. An annotation is inserted
into the
file at the beginning of each section. The annotations have to be
written using
a special character style (hidden+shadowed). For each unit, the
author can (but
not have to) provide the set of outcome and background concepts. The
format for
the outcome annotation is: (out: concept-name1, concept-name2,
...).The format
for the background annotation is: (pre: concept-name1, concept-name2,
...).
Step 3. Translation to HTML. The annotated MS Word file has to
be saved
in RTF formats and translated into an HTML file by the RTF2HTML
program
controlled by some specially designed settings. All annotations and
section
titles are translated into HTML comments which have a special format.
The
resulting "InterBook file" is just an HTML file annotated with
several kinds of
special comments.
Step 4. Parsing into LISP structure and Serving on WWW. When
the
InterBook server starts, it parses all InterBook files and builds the
list of
section frames. Each section frame contains the name and type of the
unit, its
spectrum, and its position in the original HTML file. The obtained
LISP
structure is used by InterBook to serve all the available textbook on
WWW
providing all advanced navigation and adaptation features. All
content which
the user will see on the screen is generated on-the-fly using the
knowledge
about the textbook, the user model, and HTML fragments extracted from
the
original HTML file. These features of InterBook are based on the
functionality
of he Common Lisp Hypermedia Server CL-HTTP.
As we can see, our tool seriously simplifies the design of adaptive
ET on WWW
for the authors who use the approach presented in [2 An Approach for
Developing
Adaptive Electronic Textbooks]. It provides full support in
preparation and
serving an ET for the authors who know only how to use the MS Word
text
processor. An advanced used who have some knowledge on HTML and LISP
programming can use our tool more flexibly. For example, an author
can bypass
step 1 and 2 by preparing the textbook directly in HTML format with
annotations
provided as specially formatted comments. The author can also replace
server
response functions and HTML generating functions to implement
different
structure and different "look and feel" of the be requested by a
unique URL. To
enable the server to respond to a particular URL, this URL has to be
associated
to a response function implemented in LISP which has to generate an
HTML page
on the fly as an adaptive response. CL-HTTP includes a set of LISP
functions
for generating pages.
InterBook is expected to be used with Netscape 2.0 or 3.0 browsers.
It uses
advanced features of these browsers such as multiple windows and
frames to
provide the user with useful and powerful interface. Main windows
used by
InterBook are the textbook window and the glossary window.
The Glossary window is used to view the glossary. The upper part of
the window
is a list of glossary concepts. The lower part of this window is used
to show
the glossary entry for a concept. For each concept the system
presents the
concept description (if provided by the author), the list of section
titles
(selected from all available textbooks) which present the concept
(i.e., which
have it as an outcome concept) and the list of section titles which
require
this concept (i.e., which have it as a background concept). Section
titles are
clickable links which makes the corresponding section to be loaded to
the
Textbook window.
The Textbook window is the most important window in InterBook
interface. This
window is designed to view the main content of a textbook, section by
section.
It is divided into frames performing different functions. Main frame
of the
Textbook window is the Text window. This window shows a particular
section of
the textbook which is called current section. For a terminal section
the Text
window shows the title of the section and the section itself. For a
high-level
section the Text window shows the title, the section preface (if
existing) and
the full table of content for the section (i.e. list of
hierarchically
structured titles of its subsections down to terminal level). A
vertical bar to
the right of the Text window is the Concept bar. It is used to show
the
concepts related with the current section. All names of concepts on
the Concept
bar are links to the Glossary. The upper part of the Textbook window
hosts the
navigation center and the toolbox. The navigation center shows the
position of
the current section in the textbook: it lists the titles of all
direct
predecessors (father, grandfather, etc.) and all brothers of the
current
section. All names of the sections are clickable links. The
navigation center
serves for both orientation and navigation. The toolbox provides a
set of
buttons which are used to call additional windows (such as content
window,
search window, and prerequisite-based help window) which provides
additional
functionality.
Figure 1: Textook and Glossary windows of InetrBook
The main idea behind our tool is using concept-based indexing
to make
conventional educational material more intelligent and flexible. The
idea of
indexing is to provide the information about the content of each unit
of
conventional educational material by indexing this unit with related
domain
concepts. Previously, indexing was applied in three authoring
contexts: CAI
context, hypermedia authoring context and ITS authoring context.
Indexing was originally suggested in CAI context by Osin [1976] who
suggested a
framework for indexing CAI frames by a set of topics which it covers.
Such
indexed sets of frames were not related to any pre-scribed order of
presentation. They can be accumulated, stored in special libraries,
and re-used
by different authors to create their own courses. In the multimedia
field, a
similar idea of a re-usable database of multimedia learning material
indexed by
topics and keywords is elaborated by Olimpo et al. [1990].
Later indexing was applied in hypermedia and ITS authoring area. In
the
hypermedia authoring area, an idea of indexing was elaborated by
Mayes, Kibby
and Watson [1988] in the StrathTutor system. They stressed additional
preference of indexing the frames of learning material - the
possibility to
indicate related pairs of frames not by tedious glossary linking of
pieces of
learning material together, but dynamically, on the basis of
similarity of
corresponded sets of topics. In the ITS authoring area, indexing was
applied to
turn traditional CAI into a "slightly intelligent" ICAI [Elsom-Cook
&
O'Malley 1990; Grandbastien & Gavignet 1994; Vassileva 1992].
"Slightly
intelligent" ICAI are based on both the CAI and ITS paradigms. The
teaching
material is not generated as in 'orthodox' ITS, but stored in
CAI-like frames.
However, these frames are indexed with the concepts from an explicit
domain
model network, so they can be selected intelligently. The most recent
application of indexing on the crossroads of the above directions is
hypermedia-based ITS which use indexing technology to connect the
learning
material represented in a hypermedia form with the domain knowledge
base: SHIVA
[Zeiliger 1993], ITEM/PG and ISIS-Tutor [Brusilovsky, Pesin &
Zyryanov
1993].
Indexing shows to be a relatively cheap and useful technology for
authoring
"more intelligent" hypermedia and CAI systems. We argue that it is
the relevant
technology for developing more adaptive and intelligent educational
applications of WWW. Currently, we can name only ELM-ART
[Brusilovsky, Schwarz
& Weber 1996] as an example of a WWW adaptive electronic textbook
based on
indexing. We expect that the WWW will boost the research and
development work
on adaptive electronic textbooks. We hope that our tool will be
useful for
those authors who are interested to make their ET adaptive and to
serve it on
WWW.
[Brusilovsky 1995] Brusilovsky, P. (1995). Adaptive learning with
WWW: The Moscow State University Project. In Held, P., &
Kugemann, W.F. (Ed.), Telematics for Education and Training .
Amsterdam: IOS. 252-255.
[Brusilovsky 1996] Brusilovsky, P. (1996). Methods and techniques of
adaptive hypermedia. User Modeling and User-Adapted Interaction, 6
(2).
[Brusilovsky, Pesin & Zyryanov 1993] Brusilovsky, P., Pesin, L.,
& Zyryanov, M. (1993). Towards an adaptive hypermedia component
for an intelligent learning environment. In Bass, L.J., Gornostaev,
J., & Unger, C. (Ed.), Human-Computer Interaction (Lecture Notes
in Computer Science, Vol. 753). Berlin: Springer-Verlag. 348-358.
[Brusilovsky, Schwarz & Weber 1996] Brusilovsky, P., Schwarz, E.,
& Weber, G. (1996). ELM-ART: An intelligent tutoring system on
World Wide Web. In Frasson, C., Gauthier, G., & Lesgold, A.
(Ed.), Intelligent Tutoring Systems (Lecture Notes in Computer
Science, Vol. 1086). Berlin: Springer Verlag. 261-269.
[Elsom-Cook & O'Malley 1990] Elsom-Cook, M.T., & O'Malley, C.
(1990). ECAL: Bridging the gap between CAL and intelligent tutoring
systems. Computers and Education, 15 (1), 69-81.
[Goldberg, Salari & Swoboda 1996] Goldberg, M.W., Salari, S.,
& Swoboda, P. (1996). World Wide Web - course tool: An
environment for building WWW-based courses. Fifth International World
Wide Web Conference.
[Grandbastien & Gavignet 1994] Grandbastien, M., & Gavignet,
E. (1994). ESCA: An environment to design and instantiate learning
material. In de Jong, T., & Sarti, L. (Eds.), Design and
production of multimedia and simulation-based learning material.
Dordrecht: Kluwer Academic Publishers. 31-44.
[Kay & Kummerfeld 1994] Kay, J., & Kummerfeld, R.J. (1994).
An Individualised Course for the C Programming Language. Second
International WWW Conference "Mosaic and the Web".
[Lai, Chen & Yuan 1995] Lai, M.-C., Chen, B.-H., & Yuan,
S.-M. (1995). Toward a new educational environment. 4th International
World Wide Web Conference.
[Mayes, Kibby & Watson 1988] Mayes, J.T., Kibby, M.R., &
Watson, H. (1988). StrathTutor: The development and evaluation of a
learning-by-browsing on the Macintosh. Computers and Education, 12
(1), 221-229.
[Nakabayashi et al. 1995] Nakabayashi, K. et al. (1995). An
intelligent tutoring system on World-Wide Web: Towards an integrated
learning environment on a distributed hypermedia. ED-MEDIA'95 - World
conference on educational multimedia and hypermedia, AACE. 488-493.
[Olimpo et al. 1990] Olimpo, G. et al. (1990). On the concept of
database of multimedia learning material. World Conference on
Computers and Education, North Holland. 431-436.
[Osin 1976] Osin, L. (1976). SMITH: How to produce CAI course without
programming. International Journal on the Man-Machine Studies, 8,
207-241.
[Schwarz, Brusilovsky & Weber 1996] Schwarz, E., Brusilovsky, P., & Weber, G. (1996).
World-wide intelligent textbooks. ED-TELECOM'96 - World Conference
on Educational Telecommunications, AACE. 302-307.
[Thimbleby 1996] Thimbleby, H. (1996). Systematic web authoring.
Symposium "The Missing Link: Hypermedia Usability Research & The
Web".
[Vassileva 1992] Vassileva, J. (1992). Dynamic CAL-courseware
generation within an ITS-shell architecture. 4th International
Conference, ICCAL'92, Springer-Verlag. 581-591.
[Wenger 1987] Wenger, E. (1987). Artificial intelligence and tutoring
systems. Computational approaches to the communication of knowledge.
Los Altos: Morgan Kaufmann.
[Zeiliger 1993] Zeiliger, R. (1993). Adaptive testing: contribution
of the SHIVA model. In Leclercq, D., & Bruno, J. (Eds.), Item
banking: Interactive testing and self-assessment. Berlin:
Springer-Verlag. 54-65.
Part of this work is supported by a Grant from "Alexander von
Humboldt
Foundation" to the first author and by a Grant from "Stiftung
Rheinland-Pfalz
für Innovation" to the third author.