Towards Integrated Learning and Teaching Environments

(position paper for the WWW 6 Workshop "Teaching and Learning with the WWW: Learning from the past")

Name: Daniel Schneider and Patrick Jermann
Affiliation: Tecfa, Faculté de psychologie et des sciences de l'éducation, Université de Genève
Address: 9 route de Drize, CH-1227 Carouge and and


Issues for discussion

We would like to discuss several groups of issues in this workshop:
  1. As first issue we'd like to roundup opinions of the participants about the state of "Virtual Campuses" and more generally "Education on the Internet". This is partly related to our well founded hope that after 2 years of near absence we will manage to save our WWW Virtual Library on Educational Technology from "near death". A first trial is here (might be moved again). It's based on a MSQL data-base that needs to be filled in. In addition our unit has built a Swiss server for education (German/French only so far) that could profit from some insights.

  2. As second issue we would like to address dynamic, persistent, multi-user worlds and their role in education (also in relation to more groupware-oriented "Virtual Campuses"). Such environments will without doubt become quite popular although it's a bit difficult to see in which direction we are heading and how future "educational worlds" might look like. We do have our own text-based multi-user world (the TecfaMOO) and currently are investigating VRML-based technology.

  3. As third issue we wonder how we could better integrate research in advanced learning environments with the Internet. We'd like to refer to work done in the "Artificial Intelligence and Education" (AI&Ed) Community which over the last years tried to combine principles derived from artificial intelligence, cognitive science, human-computer interaction, and related fields. for the design of systems to support learning.
We are aware that some of these issues do overlap and that we ask too many and sometimes too general questions. Let's see what we can do within a one-day workshop.

1. Virtual Campuses: What can we do ?

In discussing the role of technological support in education, Sandberg, (1994: 225) identifies the components of a (technologically rich) learning environment. These components must all be there in order to optimize learning. However, they can be ``implemented'' in many different ways. Each component has functionalities that we should insure in a "virtual Campus": Sandberg's typology in GIF
  1. ``Teacher'' component: Its role is to provide something between loose guidance and direct instruction. It can be a human agent (present or distant), an intelligent agent, instructions like some text books provide, etc. This component provides information from the syllabus to the task level.
  2. ``Monitor component: Ensures that something is learned. A role taken by either the human teacher, the learner (self-control) or by some program.
  3. ``Fellow learners'' component: Improves the learning process (some research tries to implement artificial ones).
  4. ``Learning material'': Contains what has to be learned in a very broad sense (knowing what, knowing how). It can be computational in various ways (exploratory hypertext, lesson and task oriented hypertext, simulation software, task solving environments, etc.).
  5. ``External information sources'': All kinds of information which is not directly stored in the learning material (e.g. additional material, handbooks, manuals, etc.).
  6. ``Tools'': Everything which may help the learning process other then the learning material (e.g. calculators, communication software, etc.)
  7. ``School'' [a category we added]: Something that provides a curriculum and does student administration.
At a different level, we can identify the following three key components: (1) Information, (2) Communication, and (3) Computer-based learning

They typically come in separate form, Information and Communication being pre-dominant in the Internet World.

Issue 1 Questions:

  1. Do you agree with Sandberg's figure and can we fill in every "function" with a good example (articles, deployed systems) ?
  2. Now how can we combine Information, Communication and Computer based Learning systems to gain additional synergy ? After all, one major advantage of Internet is the possibility of integrating different kinds of content and communication.
  3. What is the state of CBL over the Internet ? Typically, most CBL programs are built with specialized tools such as (Aysmetrix's Toolbook or Macromedia's Authorware). What is the state of tools and concepts available for Internet Content providers (including plug-ins provided for traditional authoring tools) ?
  4. The major part of educational material on the WWW are course notes. Next, we find some structured and well designed teaching materials including texts, exercises and tests. More sophisticated applications are rather scarce, although educational philosophy advocated is changing as shown for example at the CyberEd site . Will we see more interactive "learning-by-doing" on the WWW in the nearer future ? And is it necessary ?
Here are some resource pages referring to learning and teaching theories: Question: Any other major sites on that topic ?

2. Dynamic Worlds for learning and Teaching

Right now there seem to be 2 major strings of research and development:
  1. Educational groupware:
    On-line materials + tests+ email + forums + student management + voice/video, e.g. systems like WebCT (, the Clyde Virtual University ( and WEST (
  2. Multi-user Worlds:
    Multi-user Worlds favor social immersion within a real-time multi-user environment. Most allow user-built extensions and are built on persistent objects. For a MUD-biased index, see our Educational VR page ( Within these we find 2 major types:
A lot of design issues could be addressed. We single out just a few (for others see our position paper for the WWW 5 workshop on Virtual Environments and the WWW):

Issue 2a: Do different kinds of "educational worlds" favor different pedagogies ?

In the appendix we present a short comparison between groupware, VR and MUD - oriented worlds that answers this question positively. Issue 2b: Do different types of immersion favor different types of learning ? According to Zeltzer/ we can classify a virtual environment with the following dimensions:

  1. Autonomy: how things react.
  2. Interaction: how things are manipulated.
  3. Presence: immersion of the user.
According to Whitelock, Brna and Holland, different degrees of immersion are probably better suited for different types of learning:

lowconcept learning?
highprocedure learning

When is immersion useful and when isn't it ?

Issue 2c: How can we optimally use different media ? CMC technology implements an amazing amount of different media that can be categorized in various way (for an attempt inspired by Gay's and Lentini's "Use of Communication Resources in a Networked Collaborative Design Environment" paper, see the appendix) One of our research projects has shown that building shared problem representations (grounding) are facilitated when different media are available, but also that different users have different preferences. Different media bring different resources and constraints on grounding as well as having different associated costs. Any comments on that ?

Issue 2d: Structured vs. unstructured interfaces One of our research projects has shown that content type of utterances is related to the interface type usage. When do constrained (structured) interfaces help learners, when do they rather hinder it ? At at more conceptual level we can ask how we can "encrypt" optional conditions of collaborative learning within a virtual environment.

Issue 2e. Is it true that most things seen on the Web favor quite a traditional approach ? And is the situation really different in multi-user "worlds" ? Does the "virtual community effect" have any impact on learning, i.e. does "being there" help, at least by providing motivation ?

3. Advanced Learning Environments over the Internet ?

A lot (if not most) Advanced Learning Environments research has been addressed by AI&Ed researchers. E.g. see the goals of the AI-Ed society ( or the topics addressed in the next AI-Ed conference ( Obviously, Internet will change the Distance Teaching Market (and further enable "On the Spot" and "Just in Time Learning"), but this is not the (btw important) issue we want to address here. We'd like to discuss what's really exciting about new Internet technologies for people working on advanced learning environments. Here are (a few) of our questions:
  1. Nothing will change in the immediate future ? After all we did have networked computers for years in education (and nicely written text books too) and the issues at hand have not become easier. Proof for that statement is that most Internet educators ignore advanced research in learning technology (and the other way round). Have a look at the sessions of the last European Conference on AI in Education last fall. (or become a reader of the International Journal of Artificial Intelligence in Education). Also interesting are books like Tiffin's and Rajasingham's "In Search of The Virtual Class" who assign a major role to AI but blatantly ignore 15 years of well organized research in that area. Exceptions are rare, e.g. Brusilovsky's work.
  2. Can we make learning environments more fail-safe by adding humans, e.g. tutors and co-learners ? (Artificial Tutors tend to get stuck in some situations.) In addition, since many researchers in advanced learning environments point out to "cooperation" as the hot topic, Internet may have it's chance there by providing at least better Collaborative Interactive Hypermedia on a wider basis.
  3. Designing and prototyping software is very expensive. Over the Internet, cooperative projects could be done more easily, although ideal instruments are still lacking in our opinion. The "all-in-one-potential" of "Cyberdesktops" also does have appeal for development as it is shown for example in some MOO servers that are used for all sorts of purposes (virtual communities, on-line classes, simulations, collaboration research and more). In such worlds, people can built on top of what others are doing. Could we design a "virtual place" where researchers could develop their projects based on some shared infrastructure ?
  4. Virtual Agents: In the AI-Ed community agents refer to tutors, coaches, co-learners and such whereas in the Internet community they mostly name some kind of smart search engines. Where is the potential of more versatile agents on the Internet ? Also, can we build simpler agents in virtual worlds ?
  5. Persistence of user-created objects: In stand-alone applications, most things built by the user are lost at some point (or at least isolated from co-learners). Virtual worlds could boost co-learning by sharing objects with others. In addition, what could be gained from "walls with memory" in Virtual Worlds ?.
  6. How could AI help designing immersive "conceptual spaces" or "living" representations of educational content ? After all, representation is one of the key research areas in artificial intelligence and so far only smart search engines have profited from this body of knowledge. How can we represent abstract concepts or writings within a virtual space and what does it buy us ?
  7. Learning is experience: Does Internet favor that ? Can we build spaces where people experience a wide area of topics (e.g besides circular themes such as learning how to program within a multi-user programmable environment?) Furthermore: How can we avoid the "video game effect", i.e. people learning how to manipulate the "virtual" without learning from it for the real ?
  8. How could we use smart WWW servers ?. Can we get more educational value out of specialized Web servers e.g. "webbed MOO servers" and the Common Lisp Hypermedia Server ? Smart (e.g. adaptive) hypertext certainly falls into this category.

About ourselves

Daniel K. Schneider is a research associate at the Educational Technology Unit (TECFA), Faculty of Psychology and Education, University of Geneva, Switzerland. His current research interest include computer mediated communication and information tools on the Internet. He currently teaches a CMC class and a "cyberspace" programming lab and manages the computational infrastructure of TECFA. In the past he has worked on intelligent learning environments (AI & education) and cognitive modeling. He is a trained political scientist and wrote a Ph.D on modeling political decision makers.

Patrick Jermann is a research assistant at the Educational Technology Unit (TECFA), Faculty of Psychology and Education, University of Geneva, Switzerland. His current work focuses on the use of semi-structured interfaces in collaborative problem-solving. He is administrating a server dedicated to swiss educational research ( and collaborates to a programming course for educators and psychologists.

TECFA ("Technologies de Formation et Apprentissage") is a small academic unit active in the field of educational technology. It belongs to the School of Psychology and Education ("Faculté de Psychologie et des Sciences de l'Education) of the University of Geneva. TECFA's research covers a large area of interests, including: cognitive issues in learning technology, applications of artificial intelligence to education, cognitive effects of educational software, computer mediated communication and information systems in education, multimedia courseware, and distance education. Since Fall 1994 TECFA offers a postgraduate diploma (DES) in educational technology ("Sciences et Technologies de l'Apprentissage et de la Formation") within which we both use and teach Internet-based technology.


More on Issue 1

Let's look at some of the more popular www/internet technologies. Are some technologies better suited for some functions ? E.g. does the following table make sense ?

Internet in a global learning environment
What Technology"school"teachermonitorfellow learnerslearning materialExt. info sourcestools
Simple WWW**- *******
WWW with server-side scripts***************
WWW & local clients *(*)-*** ***
WWW mobile code  *******
WWW "exotic" Plugins  (*)(*)*******
Email ****** ** 
Combined technologies
(News, HyperNews)
Electronic Classrooms *******   
Virtual Campuses
(groupware and materials)
(shared applications)
   *****(*) ***
virtual worlds

More on issue 2

More on Issue 2a: Differences between VR, MUDS and Virtual Classrooms

Does the following table make sense ?

VR vs. MUDS vs. "Virtual Classrooms"
  immersive VR"Virtual classroom""MUDs"WWW pages
produced byComputer people and artistsInstructional specialistseverybodyeverybody
typical educational userskidsstudentslarger audience
mostly students
favorite subject visualizationallhuman sciences all
favored activityobject manipulationquestions/answersdiscussion /constructionlecture
(constructivist learning)
Instruction (information-do-test loopsVygotsky (proximal learning)?
deployed in education?noyesyesyes
persistant objectssometimesnoyesno
user extensibilitysometimesnoyesyes
mediainterfaces RVgroupware tools
and voice video
text, but also HTML, VRML, GUIs 
virtual presence
(via an avatar)
(virtual self)
N users2 (10)20 (100)200 (1000) 
ease of usemediumyesmediumyes
interactive 3D graphics
immersion effects
pedagogical effectivenesscollaboration and social learning
on-line textuality

More on issue 2c

Here is a short attempt list various criteria for classifying CMC media.

Media Aspects
Co-presenceAre participants "present" in the same location ?
VisibilityAre participants visible ?
AudibilityCan people talk and listen ?
co-temporalityDo people communicate at the same time ?
sequencialityHow fast do we get a reaction usually ?
reviewabilityCan people review what has been uttered ?
revisabilityCan people revise before they send something ?
archivabilityCan the result be stored easily ?
user numbersHow many users can participate at the same time ? ?
proxemicsAt what distance can people be within a medium
simultaneityCan people "utter" at the very same time ? turn-taking ?
multi-channelsHow many channels can participants use ?
multi-mediaWhat kind of media does it support ?
instrusivenessHow intrusive is the tool ?
structurednessHow does the tool pre-structure conversation ?

Each of these features appear with different weight in various "Systems":

Media Aspects of a few types
 face to
user numbersfew+55+57+10+10+5
structuredness+++++++ ++++
(This table is filled in a bit randomly for now)

What makes a difference and which difference? Does it make sense to constrain an educational virtual world to a small subset of these ?