Ulrik Schroeder, Michael Brunner
TH Darmstadt
Wed Apr 27 19:43:21 MET DST 1994
We exploit HTML and Mosaic to construct a local and non-generic hypertext version of a lecture manuscript ``Basics of Computer Science'' at Technical University of Darmstadt. So far, we do not intend to implement a virtual university like for instance GNA or Paieida, which use Internet as a campus-like base for distant education.
There are many advantages in utilizing hypertext technology to prepare a lecture's manuscript:
This results in better motivation of students and easier understanding of the presented material. But most important, students can actively take part in their education by interacting with the content of the script, and with the staff or each other. Active reading means
How an active script can be realized will be discussed in section 4. Another advantage of hypertext is the possibility to implement variant views on the presented material. The lecture view presents the material as tutorials on several topics, usually starting at a coarse level (introduction), giving basics on different topics, before going into detail with advanced stuff based on the previously taught lessons. A modular view concentrates on a specific topic without giving background information on other topics. This view can be used while repeating lessons and looking up information. A compressed version might only give formulas, axioms and proofs, to be used as collection of formula, while taking an examination. The example and exercise view sums up animated examples and exercises. Preparing lecture material for variant views takes 3 steps
Views could be implemented taking the Intermedia approach of separating
nodes and webs, using generic or procedural links. But, how can views be
established using Mosaic, HTTPD, and HTML, when links are integrated
within the contents of nodes? One simple approach is generating different
kinds of tables of contents, leading to different sequences and may be some
different nodes. But after following the initial link into the common web of
information nodes further links can not be implemented view specific. We
would like to use conditional links leading to different places
depending on the path followed to the current node. At the moment,
conditional links could be implemented as interactive forms which e.g.
inquire about the students experience and activate another Mosaic process
with the ``best fitting'' node. A better solution would implement conditional links in HTML, which could look like this:
<A if condition then HREF=URL else HREF=URL> text </A> or
<ASWITCH expression
case value1: HREF=URL
case value2: HREF=URL> text </ASWITCH>
Attribute values should be set while following links (e.g. the starting
points from view specific table of contents) and be evaluated while
calculating the URLs of the current node's links. In our opinion, a
view should give navigation proposals (sorted links by suggestion priority)
but should not prevent the user from following links of different paths.
Tools automatically following links (e.g. printing a complete document)
should then use the next-link with highest priority. HTML browsers
should sort links by priority or highlight more important links by font
and/or color. Another enhancement to HTML would either implement a list
of rated links:
<ALIST> ATTR=a HREF=URL TEXT=, ATTR=b HREF=URL TEXT=, </ALIST>
The advantage of developing a script on a modular base by concentrating on a specific topic and giving background or interrelated information as links can be compared to a modular (structured) programming style. But general information structure is more complicated and can not be structured in a way as programming modules, because in general one will not be able to define interfaces between information units. This is why a didactic concept must be validated in practice, by evaluating statistics and interviewing participating people (students and staff).
Tracing the usage of the document is one way to improve the organisation of a lecture and the design of its corresponding manuscript by evaluating
Statistics inform about
Even if hypertext scripts have many advantages over conventional forms, there are still some drawbacks considering orientation. Many hypertext systems support orientation with graphical or textual (smalltalk-like) browsers, some of them utilizing a fisheye view, marking the current node and may be the path, which has been taken to reach the node. Mosaic does not implement a graphical browser, because drawing a graph of nodes in the Internet does not make sense with its complexity. Orientation supported by Mosaic are its
To enhance user orientation we would like to introduce special user annotations, which
One problem is to control proper authorization: Text and designed reference structure of the original manuscript must not be changed by students (except for public annotations, which should be readable to anyone who decides to follow this link). On the other hand personal annotation are necessary to model ususal learning strategies and insert own links.
Active elements are one way to increase learners motivation and receptivity by supporting explorative learning. Following the premis, that own experiences can be better remembered than passively received ones, we are going to enforce the learners initiative to work out the actual topic, in the sense of ''Learning by Doing''.
We do directly integrate exercises and examples into our document. Exercises can be processed by following a link to PSG, our interactive programming environment and interpreter for probably incomplete programming fragments, that can be tested with user input data. In this way we can enhance the readers understanding of the formal specification when he is observing the program's behaviour and can draw his own conclusions on the formal specification.
Furthermore we will integrate educational animation and debugging tools, developed at our Institute :
Besides the integration of programming environments and visualisation tools we will connect a theorem proofer to support the reader with formal program proofing.
We see our document as a platform for discussion about the course. We introduce newsgroups that can be directly read from within the document and where articles on current topics can be posted. The newsgroups will be structured hierarchically and according to the topics, like :
In addition to the higher motivation of the learner we have several new possibilities for the teacher to get information about the learners state of knowledge. So we can have exercises and their workouts on-line, which makes them easier to evaluate. We link a sequence of hints on exercises, which can be followed by a student, who gets stuck while trying to find a solution. The evaluation of the usage of these hints enables the teacher to draw a conclusion on the amount of needed support. We keep the traditional form of solving and presenting exercises in groups who can suggest their solutions to be posted in a newsgroup for discussion and further improvement.
In cooperation with IPSI of GMD we analyse the use of hyperdocuments in seminars. We use an electronical blackboard (XEROX LiveBoard) which is networked with local student computers. The lecturer can extend the usual overhead slides with the various kinds of multimedia information. The consistency of the presentation material with the digest of seminar papers will be increased.