VEDLEGG 5 The Document Template as an Elearning Resource

Helge Høivik
Center for Staff and Learning Development, Oslo University College
October 2001

Elearning is commonly defined as an Internet-based endeavor. This paper^[1] is an experience-based account in 6 steps leading up to, but stopping short of such WEB-based solutions. The purpose is to substantiate that documentalism in general and document templates in particular are core elements of computer-supported learning.

1. Initial steps

The widespread practice of personal computing commenced in the early 80’s with the diffusion of low-cost machines^[2]. Later in that decade, they provided standardized support for bitmapped displays and other interface enhancements^[3]. This made it possible to work with precision on image editing and desktop publishing. In the early 90's hypertext authoring systems with support for animation, programmed interaction and on-line audio/video became available. The following examples stem from these three period.

In addition to traditional textbooks, the students received locally written and photocopied material in response to daily and weekly needs. This gray material became the target for local digitized publishing. Spreadsheets and database handling systems on the PC or on larger systems using PCs as dumb terminals provided topical examples. With the support of cut-and-paste-functions and rudimentary drawing packages, the author could compose illustrated texts that described procedures and outcomes, assignments and so on. Such materials were written or copied to a wordprocessor, edited and printed for Xeroxed distribution. Long distance access through 300 baud modems was sufficient for the eager amateur, but too expensive and cumbersome for real networking.

The students in these classes did not have a strong command of Norwegian with two important consequences. Teaching methods had to rely heavily on the pedagogical principle of learning-by-doing. Since most published material in the relevant subject areas were available in English only, - and not particularly suited for language training -, all texts were locally and digitally authored in Norwegian.

In the first year fresh material was produced and distributed to students 10-15 pages at a time. In the second year it was comprehensive enough to create finished booklets on a local offset machine. The final collection comprised approximately 1000 pages of illustrated print.

Digital tools evolve rapidly. Whereas the business cycle of early industrialism was 10-15 years and diminishing to half of that in this century, the digital product span may be as low as 3-6 months. Each year a new and improved version of a given software or hardware entity was introduced while new and cheaper products arrived to compete with the earlier entrants.

Added to this was a more deep-seated transformation. For each iteration of the version cycle, the products were delivered with more extensive help files, help functions (wizards) and on-line manuals. Why write detailed explanations when they were already at hand? Rather than see software as mere operational tools, they constitute a new kind of learning resource. Editors and other software support both learning-by-doing and learning-by-reading. These artifacts are themselves authored (programmed) artifacts and thus combine prescriptive, normative and descriptive functions. Additional descriptive and normative texts should balance and integrate well with such resources.

Given these conditions, the problem was to maintain locally authored texts up-to-date. Materials were organized according to a simple three-tiered classification scheme: theoretical treatment, exercise and tool description. Texts of the first kind change at a relatively slow pace; tool descriptions needed frequent updates and integrated well with help-files etc., while exercises held middle ground. This approach resulted in a small number of printed books and booklets of 100-300 pages each.

With a completely digitized collection of materials and better display technology, the stage was set for a reversal. Why not use digital display and distribution technologies to replace paper as an intermediary?

If so, materials had to be reformatted to support individual on-line study, on-line presentations in class as well as paper printouts.

2. What's in a name?

Asymetrix’ ToolBook is a visual, object oriented hypertext and programming environment. The manual stated that

ToolBook uses the metaphor of a book as the basis for your application… A ToolBook application consists of one or more DOS files called books. Like a printed book, a book in ToolBook is divided into pages, which represent the application’s screens. You view the pages in windows called viewers.

OpenScript is the ToolBook programming language…that includes commands to accomplish a wide variety of tasks, from creating and managing new objects to linking functions in Windows DLLs.. The programs are compiled using the ToolBook compiler, and ToolBook runs the finished scripts at Reader level as part of your application.

In the parlance of this particular authoring tool, documents are referred to as books. But what’s in a name? To clarify the issue we might want to compare with common definitions of traditional books.

Webster’s dictionary defines book as set of written sheets of skin or paper or tablets of wood or ivory. Now,- some children’s books are made of plastic and the early versions of The Book came on scrolls. Such books evidently fall outside this enumerative and materialistic definition. A functional delimitation in Webster’s states that a book (might be) something that yields knowledge or understanding. But we assume that parents and peers, teachers and travels do precisely that. We might abstract from of the materiality of the book, e.g. describe it as a substrate that is formed or processed in certain ways (bound, folded or rolled). Added to this must be a description of how this substrate is meaningfully engraved or imprinted with marks, etc., leading us to further define engrave, (im)print, mark, sign and meaning. For statistical purposes, a book is often considered to be a non-periodical publication of 49 pages or more, which leads us to further define publishing and page and so on.

This exemplifies how enumerative and functional definitions may be deconstructed. There is a point worth remembering here and aptly worded by Wittgenstein:

Meaning cannot be found in the reductionary analysis of systems at the level of fundamental propositions. Rather, meaning is a shifting and intensely personal phenomenon contingent upon a holistic and simultaneous consideration of the whole complex of factors related to a particular topic of human interest^[4].

Looking at Webster’s definition in its entirety, which is an inventory of terminological usage, we find a good illustration of this:

Book: set of written sheets of skin or paper or tablets of wood or ivory, set of written, printed, or blank sheets bound together into a volume, long written or printed literary composition, major division of a treatise or literary work, record of a business's financial transactions or financial condition -- often used in plural (the books show a profit), BIBLE, something that yields knowledge or understanding (the great book of nature, her face was an open book), the total available knowledge and experience that can be brought to bear on a task or problem (tried every trick in the book, the book on him is that he can't hit a curveball), standards or authority relevant in a situation (run by the book), all the charges that can be made against an accused person (threw the book at him), position from which one must answer for certain acts, ACCOUNT (bring criminals to book), LIBRETTO, script of a play, arrangements for a musician or dance orchestra, musical repertory, packet of items bound together like a book (a book of stamps, a book of matches)

If the Sausurrian signifié is so elusive and divergent, so dependent on the techniques of print, so linked to the practices of writing, accounting, standard setting, religion^[4] and law etc., we might expect a similar plenitude in the digital forms. This is indeed the case.

Apart for the bookishness of the Ebook, PDF and Word formats, the most important metaphors for digital textual structures are slide show, flowchart, card stack, object hierarchy, cast/score and frame/link^[6]. System functionality is also enhanced by generic scripting languages, e.g. interpreted or semi-compiled programming languages. Let us look closer at each of these classes:

• As the name implies, slide show systems organize material as a sequence of individual slides that contain multimedia elements. Several commercial systems are available and are widely used wherever the relevant display technologies are available. Examples are Powerpoint, Freelance Graphics, Harvard Graphics and RealSlideshow.
• The building blocks in flow control systems are display or interactive elements and a set of logical constructs. The latter are variants over the set of basic procedural programming constructs, i.e. sequencing, conditional branching and iteration. Sequencing means that displays or events follow each other sequentially. Branching occurs at sequence crossroads. Based on a given condition, the document handling system will choose one of two or more possible next steps. The test criteria may be satisfied by system-derived features, e.g. the lapse of a certain time period, or by user interaction at this or previous steps. Iteration, or repetition, occurs when the document handling system backtracks to a previous step, e.g. repeats a question when the user has given a wrong answer. In such systems the document designer draws a flow diagram depicting multimedia elements, interactions and sequencing logic. The prominent example here is Authorware.
• In object hierarchies each design element is treated as an object. Except terminals like simple geometrical shapes, bitmaps and alphanumeric characters, the objects may contain other objects. Some or all objects may also host program statements (scripts). They are used to describe interactions and object behavior like movement and color change. They perform traditional forms of data manipulation like arithmetic and string operations. Toolbook is one example.
• A card stack is the object hierarchy’s poor cousin. The card acts as a container of multimedia fragments while a set of cards forms a stack. Hypercard that was delivered for free with some version of the Apple computer popularized this approach.
• Cast/score systems implement the musical score or movie storyboard metaphor. The multimedia elements are placed on horizontal tracks on a timeline. Alignment across tracks represents synchronicity. Some systems also implement scripting capabilities. Such systems have achieved a certain level of standardization with development of SMIL (Synchronous Multimedia Integration Language) which is an XML Document Type Definition (DTD).
• Finally, a frame/link system is built of individual frames (files, database records) with embedded multimedia elements and predefined link constructs, - the various HTML specifications for the World Wide Web being the most widespread.

The tool used for our our productions is a variant of the object hierarchy type. An application is here composed of pages contained in books and several books may be integrated into one application. A page is the superimposition of a foreground on a background with backgrounds as the container of foregrounds. The latter may consequently be shared across pages. Both layers may contain various objects, e.g. text fields, images, geometrical figures, buttons, etc., that may be visible or invisible (hidden) at a given time. Text fields content is structurally organized as a sequence of lines composed of words composed of characters, which may be individually addressed and manipulated. The system also supports the construct of views through the use of viewers. Several viewers may be open simultaneously, displaying different (parts of) pages. Statements in a particular scripting language control interaction and flow. It its therefore possible to emulate most other architectures. All computational resources on the host system are also available from within an application, either through dynamic data exchange (DDE) or in dynamically linked libraries (DLLs).

The tool itself was complemented by a number of auxiliary programs to refine colors, create cursors, design menus, build database interfaces, etc., a substantial set of on-line reference material, an interactive tutorial and collections of ready-made books, scripts and royalty-free images, icons and other page elements. A number of other tools were also used, e.g. word processors, a database handling system, a statistical spreadsheet, sound editors, paint and drawing programs, an Internet browser and digital camcorders.

3. One Instructional Case

When an audience and a lecturer are unknown to each other, the first minute or so is important to establish mutual trust. A social contract is established between them to the effect that we, the audience, will listen to you if you say something worthwhile and I, the lecturer, will manage do so if you are an attentive audience. The responsibility for this is mutual, but overtly carried by the chair and the speaker.

In order to present main themes and generally get up to speed, the opening lecture for this particular seminar was therefore supported by a tailor-made multimedia document.

A common rhetorical approach is to manipulate the visual and the aural, - light and sound. Such techniques were used in our case. The room was darkened to the score of well-known and solemn classical music while a sequence of images were projected on the wall. They depicted artifacts from the history of writing like cave paintings, handwriting, early printing tools and printed matters, telecommunication satellites and computers. These were progressively overlaid with snapshots of students at work. Just as the shuffle of recent arrival to the seminar had settled, a recorded student voice exclaimed I never go there! (to the university library). Having thus caught the audience’s attention with a counterpoint, the remaining part of the presentation consisted in displaying each page and the various elements within the page of the document.

The document structure allowed the lecturer to move between pages and manipulate the stored elements, e.g. written text, images, sound bites, program scripts, etc. at will while communicating in the oral or written. In some cases he read aloud that which was already on display, in other cases the messages of texts and images were only expanded upon. At certain points, the presentation was suspended for short oral interactions with and within the audience. As evidenced by the confusion that several participants expressed later during the seminar, the presentation did seemingly give a movie-like quality to the presentation. Had they seen a film or a slide-show? I will explain in some detail how such effects came about.

Fig 1. The upper pane of page two of the document progressively showed photographs of the students as they were interviewed. For each new respondent, the picture of the previous one was displaced one position to the left.	Fig 2. Written transcriptions of each answer were reproduced – as they were offered - in the lower right pane. These summary statements were then collated in the left pane with the introduction of a new person.

Page two of the document was divided in three panes and presented interviews with four students about their experiences with the university library. Each interview was divided in distinct step where each step contained a singular message. The name, affiliation and level of study of the respondent were displayed along with his or her picture. Then the written rendition of each question from the interviewer was shown followed by the corresponding sound track of answers from the respondents. A transcript of this answer was also displayed. Other pages contained variations on this rhetorical theme, e.g. an interview with one graduate student about his experiences with the library as well as discussions with several librarians. To round off a separate page reproduced in writing a summary of all previous answers. With this as a referential background, the document cum lecturer now projected two questions to the audience about current and planned patterns of library use by the university students. At this point the presentation was interrupted by a group discussion.

Fig 3. Exformation used to present statistical material and its interpretation.

The technique of collating previous elements and keep them in view made the document - as documents go - a particular kind of aide-memoir for both lecturer and the audience, providing a track record of the interviews. The document thus took on the role as multilayered whiteboard. All in all, these 8 pages of the document were designed to convey a more intimate experience of the topic through individual statements in writing and speech than is possible with a standard lecture.

Exformation designates the compression of messages where we gain overview and ease of remembrance at the cost of losing detail. Statistical averages and other distribution measures are common examples. The organizers of this seminar were well aware of the need to raise questions about library roles and the current, and possible future, distribution of responsibilities between faculty, librarians, university librarians and support staff. In order to do so in a non-obtrusive way, one page was built to support such progressive exformation. It contained an abridged report of an opinion survey among library school students. The page reproduced an excerpt of a (computerized) questionnaire where students had expressed their ideas about librarianship roles, the resulting data matrix and the statistically computed correlation. The latter was represented as a live tree-structured diagram (dendrogram). By a point-and-click function, a particular part of the graph could be highlighted. The corresponding interpretation was displayed in a separate pane.

The intention was not to discuss this particular analysis, but rather to instigate discussion through analogy. A secondary purpose was to demonstrate how digital texts are useful as a vehicle to present educational, - in this case statistical -, material.

The final page covered two topics. The first reproduced theoretical statements about the relationship between material structures and social meaning. These statements were scanned from a printed treatise on sociology. But they were typographically altered for reproduction. The stroke was enlarged and rendered in white on a black background to emphasize theoretical weight. This is but one small example of how digital technology enable authors to collect, duplicate, reassemble and disperse texts. Let us designate such processes in their entirety as copy-transformation (sometimes referred to as repackaging).

The multimedia document consisted of 11 pages, but this number belies the structural complexity. They whole document comprised 132 subordinate objects as well as 500 unoptimized program statements that were dispersed over 50 separate scripts. It required 1.3 MB storage space with an additional 5 MB for uncompressed video and sound. Except for the first page, where the play and display of sound/images was preprogrammed, each step in the presentation and the traversal of pages was under lecturer control.

We may describe the production process as follows:

A multimedia document conveys different sensory elements. In this case of a lecture-oriented document they comprised recorded music and interviews, photographs, commercial clipart, a screen-captured movie, a newspaper clipping, citations from a book on sociology and statistical material at different levels of exformation in numerical and graphical form as well as texts and images created from scratch.

Those parts that came in analogue and mediated form, e.g. music, interviews and newspaper clippings, were transcribed to digital formats by scanning equipment, OCR software, AC/DC-converter or by retyping. Interviews were recorded in analogue form while recorded music was (automatically) translated to this format from a digital recording in the CD-ROM Red Book format and then re-digitized. A digital camera captured pictures in bit-map format. These were stored and extracted as analogue video signals and re-digitized once more to bit-map files. The newspaper clipping was scanned to capture one picture while the written texts was optically recognized and manually corrected in a word-processing program. Statistical material was partly captured using cut-and-paste tools from the computer display. One part was extracted from an input form to a database handling system. Another was taken from a screen dump of the data matrix file. The third element was manually constructed using vector-based drawing tools.

Reminiscent of the cut-and-paste function from our introduction, this obviously indicates a transition from one type of document structure and one type of publishing infrastructure to another.

4 Document Templates

To tailor each digital presentation or booklet for a particular event is too costly and time-consuming. As with printed matter, there is a need for standardized containers. But first a comment on the issue of templates as such.

Entities may be considered in the passive as objects of contemplation or in the active as adjuncts to performance, - as sources, means or ends. Here we are concerned with forms and the processing of forms, e.g. order sheets and order entry; library catalogue cards and cataloguing; application forms and procedures for entry to a country or an institution of higher learning.

In the passive a paper form may be considered to be a document template with well-defined semantics. Related performances, i.e. the active form-oriented tasks, consist of steps that string together in specific ways. In the abstract, these patterns of behavior may be represented by formulas, - commonly referred to as business rules -, that are directly applied to the form and, in this capacity, indirectly to the objects that are described by the form. They are links in a highly structured chain of divided labor. Let me illustrate this with the procedure for charging books in an academic library in the 1930's:

… The borrower filled out his or her name, address, and student number on a call slip, along with the author, title, and call number of the item to be charged. The borrower presented the slip at the circulation desk. The slip was then sent to a library assistant in the closed stacks who retrieved the item and returned it and the call slip to the circulation desk. The desk attendant stamped the due date in the book and on the call slip. The book was issued to the patron and the call slip dropped into a box for later filing into the main circulation file^[7].

The meaning of the fields and contents of a form must be understood in context as the inner relationship between form and formulae, - normally by observing the kind of organizational structure in which the form is used. The citation above refer to a social province consisting of borrowers and lenders (librarians and library assistants instituted as the library) in the context of formal studies, the reciprocal activity of lending/borrowing, e.g. circulation, and socio-material entities like books stored in stacks, circulation desks and other operational units. We may thus say that form-oriented tasks are circumscribed by complementary social relationships or roles, e.g. library patron/document, consumer/producer, government/taxpayer, etc.

Processing forms is conducted as if each instance is similar to all others except for a limited set of measurable attribute values, i.e. the instance is treated as a member of a mathematical set^[8]. This works well enough when we deal with simple identifiers and quantifiers, e.g. order or pay for mass-produced goods. In librarianship the physical features of a book like number of pages, height and the (binary) absence or presence of illustrations and tables belong to this category. Disregarding the physical implementation, a relational database may be understood in this way as (parts of) organizational knowledge encoded as schemas, constraints, distribution architectures, the syntax and semantics of query languages and the design of their interfaces.

Fig 4. DYNDOC is a digital textbook template.

As we will see, our document template is a descendant of such database structures. Whereas administrative forms reflect administrative conventions, such textbook templates must reflect instructional and learning practices. The task is here to identify the social role and function of the artifact.

With this in mind, we will look closer at one document schemata, the Dynadoc template[9] written with ToolBook/OpenScript.

Dynadoc books are composed of three types of pages:

• Standard Pages (Fig 5, right) containing panes for tutorial text, illustration/ animation, exercises and a notepad.
  
• Variant Pages used to display information that does not easily fit the standard formats.
  
• System Pages comprising a Front Page (Fig 5, left) with author name and book title, several pages for administrative and authoring purposes and the Index Page (Table of Contents).

Fig 5. Front Page (left) and Standard Page (right). (1: Tutorial text 2: Image area 3: Notepad 4: Table of Contents 5: Button bar)

In traditional books, the reader use bookmarks or dog-ears to be able to return to his current position. In the Dynadoc system, the user is taken to the current position by default. But to frame this event, i.e. to give an impression of contextual coherence, an excerpt of the front page is overlaid the standard page for a short interval at startup (Fig 6.).

Fig 6. Table of Content is available as a pop-up window on the front as well as pages (left). When a book is revisited, it is opened at the last visited page. In this case and to reestablish context an excerpt of the front page is shown for a short time (right).

Fig 6 also shows (left) how the Table of Content’s viewer is overlaid the front page. This feature is available on every page. The user moves to a given page by selecting the entry in the Table of Contents frame. Optionally this frame can be kept floating above the page to ease leafing.

The screen layout is modifiable. Of particular importance is the option to enlarge the image area (Fig 7). This was implemented to enhance legibility during overhead presentations. The content here consist of one or several images, in many cases with some animated behavior.

Fig 7. Enlargement of Image Area for overhead presentation.

A similar feature is also available for tutorial text as such as shown in Fig 8.

Fig 8. Tutorial text can be enlarged for better legibility on personal computer screens and during overhead presentations.

The book contains a Notepad where students can take notes during lectures or study. The notepad area can be enlarged according to need. This is useful in particular when students want to review their previous notes (Fig 9.)

Fig 9. Clicking on the right mouse button in the Notepad area toggles between normal and enlarged display.

Fig 10. Exercise viewer. An arrow in the button bar window indicates an exercise (1). The exercise viewer is activated when clicking the button, here seen in a checked state. The exercise viewer floats above the page and has buttons for activating a Windows directory (3) and a DOS level command interpreter (4)

When a page contains exercises, this is marked on the toolbar. By clicking a button here, the exercises text is displayed in a separate viewer (Fig 10). For those cases when an exercise involves working with particular program or data files, a Windows file directory or a new DOS command interpreter can be launched from the viewer.

This feature is not strictly necessary since students may equally well launch programs and find data files using the normal task switching operations of the Windows environment. But some students are not too familiar with the operating system tools available to them. The textbook thus acts as a permanent starting point, a kind of "home page", for other chores.

Standard pages do not accommodate all required formats. The Dynadoc book therefore contains variant pages without predefined panels.

Fig 11 shows an example of an online demonstration and exercise page for working with B- and B+ tree index structures.

Here the student can select one of the two data structures and work with insert ("Sette inn") and delete ("Slette") operations on a number of predefined key values. The page can also be printed for paper-and-pencil work.

Fig 11. Variant page (slightly modified)

This structure was implemented using ToolBook "page" and "background" objects:

"ToolBook is an object-oriented environment; all the visual elements of your application - buttons, fields, graphics, viewers, and even the pages and backgrounds - are objects. Each object has a set of properties that define the object’s appearance and behavior." [from the manual]

Fig 12. A ToolBook viewer displays pages that are overlaid a background. Both backgrounds and pages can contain objects (text fields, buttons, graphic elements etc.). A background can in addition contain recordfields that corresponds to database records.

Of particular importance are ToolBook recordfields. A recordfield has unique content, but the same properties on each page and Dynadoc pages are implemented with unique alphanumeric text and images on a common background. The other page types were defined by having different backgrounds.

Fig 13. Recordfield content may be stored with the ToolBook application or in an external database.

Recordfields correspond, as the name implies, to database records and these may be linked to proper databases. It is also possible to use custom made Dynamic Link Libraries (DLL) to communicate with other systems for this purpose.

If data are stored in a dedicated database, ToolBook pages act as a visual front-end (client) in a client-server architecture. This solution is architecturally the same as if another client solution had been used, for example a Web browser. Servers may reside on a local area or global network. But them may also reside on personal computers or be delivered on compact discs.

Fig 14. A general model for access to digital textbooks. Database servers delivers material on demand to interactive client-side tools.

The client-server model was not implemented in this version of Dynadoc for two reasons.

• On a number of pages, the image area contain relatively complex graphics. At this time it was not possible to store these as binary objects in a BLOB (Binary Large Object) field in a database. They might have been stored there as program scripts, though, but this would demand all too much design work. Response time would suffer as the script would have to be interpreted and the relevant objects redrawn each time a page was visited.
  
• Some Dynadoc pages contain specialized mathematical notation. The relevant systems did not support this at the time of production.

5. Navigation.

All Toolbook objects are sequentially stored and numbered when they are created in ToolBook. When browsing a ToolBook application, pages are "turned" in this sequence. The ToolBook’s OpenScript programming language has "go next page" and "go previous page" commands. The next/previous page relative to the current page is defined by the internal sequence number. It is also possible to address a given page by its internal number.

Since I wanted more flexible navigation, the Dynadoc system contains several named and timestamped sequences.

Fig 15. Page sequences is maintained by arrays of page names.

A sequence (see Fig 15) is here defined as a series of topically coherent pages that the student may traverse. Technically it was implemented as a one-dimensional array structure containing page names. An internal counter keeps track of the current entry (current index value) for this table. I implemented a "go next page" command that increments the current page index by one. The corresponding page name is fetched from the sequence array, and a "go page <page name>" instruction is issued to the ToolBook page handling routine. For backtracking, the counter is similarly decremented by one. This allows for the construction of several selectable sequences, not necessarily containing all pages in the book.

Fig 16. Navigational options

Excerpt from the on-line help system showing the functions of the navigation key (left, marked with the icon of an open book). The left mouse button advances the reader page while the right button backtracks. The extended table of contents (shown to the right) allows the user to select a new named or dated sequence. For illustrative purposes the calendar is here overlaid the content of the based sequence (named "Physical").

Fig 17. The print function allows for selection of textbook material, assignments and student comments from the current or from all pages.

A named sequence is a structure the organize larger excerpts of the book. A timestamped sequence, selectable by clicking a date in an online calendar, contains material that is used on a particular point in time. As part of the default configuration, the system stores a physical sequence corresponding to the systems internal page numbers. Another default sequence is based on alphabetical page names. Additional sequences are authored by the teacher and stored internal to the systems or externally as ASCII-encoded text on separate files. External sequences can be imported by the user, thus allowing for new views of the material to be distributed to users as the course develops.

The exercise viewer can optionally be kept floating above pages as the users thumbs through them. This is useful for review of problems together with tutorial text and personal notes, in particular when the student has written answers to exercises in the Notepad.

Material can be transferred to a local or networked printer for paper copies. The user can select to have the current page only ("Denne side") printed as a copy of the screen image. Alternatively the fields of the document can be printed in a database "report" format. Here all possible combinations of tutorial text ("Dokument"), Exercises ("Oppgaver") and own comments can be selected.

Fig 18. Page network.

Pages can be linked to each other as an associative network (Fig 18). If such a web is defined for the title, the user may digress from a given sequence in order to traverse the network links. At any time he or she may return from this web mode to the point of departure in the previous sequence.

The system keeps a log of all pages visited during one session. It is possible to move back to any of these previously visited pages by calling forth the history list and selecting the page of interest.

Additional pages can be added to a Dynadoc textbook during a course from a separate import procedure. This was used extensively in the initial phase of content development. Students would then import pages from a network file.

6. Conclusion

This paper is an account of some desktop publishing tasks confronting this particular teacher over a period of 15 years. Simple word-processing and cut&paste-functions were initially used to produce the gray material of handouts, exercises etc. Certain social and technological conditions instigated more extensive publishing of materials, terminating in a need for learner-oriented document templates. The later may be taken as a particular kind of form that is circumscribed by complementary social roles and procedures; e.g. those relating to reading, presenting/commenting, taking notes, problemsolving etc. For their worth, these experiences indicate that the social structure of elearning environments and the document structure of textual material mutually constitute each other; e.g. that the subject area of documentalism lie at the heart of elearning design.