Generating Structured Educational Web Sites Out of Stand-Alone Applications
Denis Helic
Institute for Information Processing and Computer Supported New Media (IICM), Graz University of Technology Graz, Austria
e-mail: dhelic@iicm.edu
Nick Scherbakov
Institute for Information Processing and Computer Supported New Media (IICM), Graz University of Technology Graz, Austria
e-mail: nsherbak@iicm.edu
Abstract:
This paper tries to clarify current sitaution at the field of authoring of large educational hypermedia databases. The special attention has been given to the problems that the authors have to deal with while creating such databases. One of the most important and currently biggest problems is deficiency of an easy-to-use authoring system, which would allow the authoring of a consistent navigable structure of such sites. Authoring of a convenient navigable structure becomes more and more unclearly with the increasing number of documents to be interrelated, and especially in a situation when a number of authors work on a single project. This work offers a posible solution of this problem based on the authoring of navigable structure using an another logical linking model called HM-Data model, instead of most primitive node-link model. The authors consider that the principle of hypermedia composites, which is the main idea of HM-Data model, is more suitable for this purpose then node-link model. Further, paper briefly shows, through an example, HM-Card - the authoring system which implements HM-Data model as an easy-to-use system allowing authoring based on this new principle. Some ideas about posible new applications of this data modeling approach are presented in the conclusion chapter.
1. Introduction
A hypermedia system can be seen as a special kind of a database system which provides a unique and non-sequential method of accessing information. The essential features of so-called node-link hypermedia data model are nodes (HTML documents) and links. Documents can contain text, graphics, audio, video, animation, and images while links connect nodes related in a certain manner. It is the linking capability which allows the non-linear organisation of text. Currently, many organisations consider using hypermedia as an advanced educational media. Such educational hypermedia databases containing hundreds of documents are normally referred to as WWW Educational Sites.
The comprehension and general quality of a WWW Educational sites depends on the reader's ability to construct a coherent mental representation of the educational information. It is the author's responsibility to ensure the construction of the database as a coherent entity. The construction of a coherent hypermedia databases can be considered to be a design problem. There are no established guidelines for authoring such databases. Going beyond such well-known recommendations that a hypermedia database should consist of the following three components - the content part, the organisational part, and the presentation part.
Normally, authoring of an Educational WWW site is carried out on a local authoring site, where a big number of HTML documents are created using such an easy to use WYSIWYG authoring environment as MS Front Page or Netscape Page Composer. After creating a sufficient number of documents they are interrelated by means of computer navigable links and the whole course is uploaded into a HTTP server where it becomes available for remote access.
While problem of authoring HTML documents got a scrutiny, and there even exists a number of solutions implemented as so-called HTML Editing Systems, the problem of navigating, or more precisely, of authoring of a convenient navigable structure helping users to construct a coherent mental representation of the educational information, does not attract much attention.
Actually, when a particular author deals with creating hypermedia links (i.e. with imposing of a navigable structure on a top of a big number of HTML documents), this task is far from being a trivial one. The situation is considerably aggravated when a number of authors work on a single project (so- called collaborative authoring).
It is interesting to note that HTML authoring software generally do not use HTML tags as an authoring paradigm. Such authoring systems normally use a publishing logical model, where an author can place objects on a particular position, cut and paste fragments of arbitrary complexity, etc., At the same time, link editing still follows the most primitive node-link paradigm.
Obviously, decreasing of a interrelating complexity of a big number of HTML documents can be done via an implementation of another logical linking model. This model should provide another logical view to hundreds of documents which should be interrelated, and thus make authoring considerable easier and even error-proven.
One possible candidate for such logical model is a so-called HM-Data Model which
addressed many important issues and even has been implemented as a stand-alone
hypermedia authoring system HM-Card.
2. The HM-Data Model
Current trends in logical data modeling can be classified into two main groups:
- Data models supporting links embedded within primitive nodes (i.e. links are local in the sense that they belong to individual documents);
- Data models supporting links as fully-fledged addressable objects (i.e. links are global in the sense that they belong to the hypermedia database as a whole).
In the HM-Data Model, information is not simply stored as HTML documents which have various
references between them but rather is organized into units called structured collections
(S-collections). An S-collection can be seen as a closed environment, it contains
a number of documents and/or other S-collections as members and encapsulates a
particular navigable structure ( i.e. a number of links between members ).
Figure 1: Internal Structure of S-collections
Thus, the HM-Data Model supports a third type of linking [4], conceptually somewhere in between the local and global link models. In the HM-Data Model, links neither belong to individual nodes, nor they are globally addressable objects, but they are encapsulated within hypermedia containers called S-Collections. By definition, links cannot point outside an S-collection, but only between its members; hence S-collections represent well-defined chunks of information, which may be re-used in various contexts without concern for superfluous hyperlinks. The model compensates for the restriction of links to local contexts by introducing memberwise inclusion of hypermedia chunks and orthogonal browsing semantics to switch local context.
There are several predefined types of S-collection. An envelope has links back and forth between every combination of two members in the collection. A folder joins members into a list, with links to next and previous members. In a menu, one specific member (called the collection head ) has links to and from each of the other members. Lastly, with freelinks, authors can link the collection's members in any manner they wish (see Fig. 1) [4].
Not only do menus have collection heads, but so does every other type of S-collection. The collection head is a member designated to be seen first when the user starts browsing an S-collection. This is important with respect to the navigational mechanism in this paradigm, while thinking about a S-collection as a hypermedia composite, which is to be built into a hypermedia educational database, its head is the page, which gives the reader the information about content of the S-collection and allows the reader to enter into navigational structure of this S-collection.
These S-collections, which can be seen as separate containers in which only links
between members of that container can be seen or followed, can also themselves be
members of other S-collections and can have either other S-collections as members.
The nesting of S-collections can be arbitrarily deep. Recursive membership is possible,
i.e. an S- collection can have itself as a member or contain another S-collection
which has it as a member. Also, each collection is required to have a unique name which
is used to support URL (Uniform Resource Locator) naming scheme.
The name is used to access S-collections by means of WWW browsers and for specifying an
S-collection for reuse (i.e. having the very same collection appear in two different
S-collections and having it linked differently in each case). The reuse is possible as links
are stored separately from documents.
3. HM-Data Model as Authoring Paradigm
Thus, S-Collections serve as special containers encapsulating local links between members. It should be especially noted that a particular link in HM-Data Model is not a document component (as in a plain HTML document), such link is encapsulated within a container. S-Collections (containers) constitute a particular context for all encapsulated links and what is important, can be authored or edited independently.
From authoring point of view, S-collections can be seen as a composite building blocks, providing higher level of data abstraction. S-collections can be combined into other S-collections, reused in different contexts (i.e. as members of other S-collections), edited independently, etc. providing a sufficient flexibility and productivity of hypermedia authoring.
Let us list here a few advantages of the model:
- links are strongly context-dependent what provides a necessary homogeneity of database structured in accordance with the model
- editing of links doesn't take any efforts since they are generated automatically whenever a new member is inserted
- the model guaranties link consistency since deletion of a document causes all links related to this document to be deleted as well
HM-Card is a new authoring system which can be seen as a further development of the Model, the system provides a convenient local authoring environment based on the concept of hypermedia composites (i.e. S-collections), authors can create navigable structures simply by means of inserting HTML pages into S-collections, all the links are generated automatically, using a predefined types of such S-Collections. S-collections can be further combined into bigger blocks (S-collections encapsulating S-collections as members).
It should be especially noted, that anchors for such automatically generated links can be adjusted in accordance with an author=92s preferences using predefined navigational toolbars.
A finished stand-alone hypermedia database can be automatically uploaded
into an HTTP server using a special "uploading" procedure which "embeds" the
links imposed by the S-Collections structure directly into HTML documents.
4. Authoring Process
HM-Card is a system which implements HM-Data Model. HM-Card is the hypermedia authoring system, which allows the creation of multimedia pages, which can be further exported into HTML using HM-Card Java Export Facilities, but this discussion goes beyond the scope of this paper. The system supports authoring speaking in the mean of the HM-Data Model. The author have an easy-to-use graphic user interface, which can be used to create, edit and to export S-collection of a different types, and further to import HTML documents in such S-collections. Of course, depending on the type of a S-collection a new inserted HTML document gets, automatically, corresponding links.
Let us imagine the following situation. An author created a S-collection of type
"menu" having documents "a", "b" and "c" as its members. The head of this S-collection
was set to be document "on". Thus, we have a situation as it is shown on the Fig. 2.
Fig. 2 An Example of Authoring Process
We have three HTML documents, having automatically created links in following manner. Documents "b" and "c", representing "casual" members of the "menu" have links to its head ( document "a" ), and the document "a" representing the head of the "menu" has links to all its members, i.e. to documents "b" and "c". The document "a" as the head of this S-collection is the document which will be first accessed while browsing this S-collection.
Let us now consider the situation, when one new document has to be inserted into
this S-collection ( document "d" ). The document "d" becomes a new member of
the S-collection, which leads to the automaticly creation of following links:
the document "d" gets a link to the head of the collection, i.e. to the
document "a", and the document "a" gets a link to a new member, i.e. the
document "d" as can be seen on the Fig. 3.
Fig. 3 Automatic Structure Update
Thus, S-collection created and edited in this manner represents one hypermedia building block, which can be used and reused for the building of different educational hypermedia databases. Hence, we have a composite based authoring.
Finally, we have to export this S-collection or the database as a whole into
HTML documents which contain, physically, hyperlinks. In HM-Card this can be done
per mouse click and then we have to upload the whole site onto an HTTP server.
5. Conclusion
It is widely recognized that the Internet offers tremendous potential for education. The construction of a coherent and lucid educational WWW site is one of the most important demands on such sites. Especially considering the type of information, which such sites contain ( i.e. educational material ) and the fact that because of that readers are not very familiar with the content of educational WWW sites the authoring of such sites has a crutial importance. The construction of coherent educational WWW sites is the question of the talent of the author, to "share" his knowledge with his "audience". The authoring of an educational WWW site, which has hundreds of documents and especially when a number of authors work on a single project becomes every day a larger problem. The maintance of such sites becomes to be imposible, while for example, one simple deletion of a document causes hundreds of links to be deleted manually, or insert of a document causes another hundred documents to be edited manually now pointing to this new document. All authoring systems dealing with this materia usualy have a publishing logical model and don't deal with problem of the creation of a suitable navigable structure of hypermedia databases. We think that using HM-Data Model as a tool helping the authors in the authoring of larger educational hypermedia databases and using HM-Card as a system, which implements this data model can help the authors solve a big number of problems they are confrotated with.