Johan van Wyk and Graham Stewart

Symposium : Constructing South African Literary History

University of Essen, Germany 5-8 October 1998

Slide presentation

The CD-ROM version of the SALit Web used to illustrate this paper is the practical output of my D.Litt thesis "The implications of e-text resource development for Southern African literary studies in terms of analysis and methodology. Thesis to be supplemented with an e-text database and experimental CD-ROM". As an application and adaptation of internal architecture, organisational structures and search features of the many different electronic literary resources described and evaluated in the course of the D.Litt study, the SALit Web on CD-ROM stands on its own as an artifact embodying the substance of that discussion. For the purposes of documentation and discussion, the SALit Web has proved to be something of a moving target. Until the production of the "Beta 01" version of the CD-ROM in September 1998, the structure and contents were altered frequently and some of the experimental features now omitted from Beta 01 will be incorporated into future versions of the SALit Web. References here to "the present version" of the SALit Web are to the September 1998 Beta 01 release.

While The SALit Web is a stand-alone resource for various learning, teaching and research activities in South African literature, it is also intended as a prototype for the further development of a virtual library of rare and marginalised South African writing, and so constitutes a potentially valuable anthology of sources for a South African Literary history. The CD-ROM is any and all of the following: a bibliographic database, a virtual library, a multimedia encyclopaedia, an archive for the preservation of marginalised texts, a textual analysis tool, and a learning and teaching programme.

This paper is intended to to provide a walk-through of the SALit Web resource, highlighting its various features, and indicating its possibilities. After an introduction locating SALit in the larger networked environment that lies beyond the fairly narrow perimeter of the CD-ROM, the paper is organised along the lines of an introductory users’ manual, starting with the opening display, and leading the user through some hypothetical search and retrieval activities, and showing what may be discovered. The sheer volume of information now included in the Web - due no small part to the considerable efforts of both Prof. Johan van Wyk, Head of the Centre for the Study of Southern African Literature and Languages (CSSALL) and Rashmi Jadhunundhan, research assistant at the Centre - makes it impossible to trace anything but a few pathways through the available resources. Only by using the CD-ROM itself can the user begin to appreciate the value it may have as a useful research and learning tool.

To begin with, we start with the "big picture", examining briefly the SALit Web and its internally-linked contents within the context of a global networked cultural heritage. Planning an electronic literary resource had to take into account present developments in electronic publishing: the increasingly common practice of publishing journals, books and encyclopaedias on CD-ROM and the Internet. While allowing the reader to view a publication much as it would appear on a printed page - but now transferred to a computer display - the electronic text may be accompanied by a powerful computerised search apparatus that delivers easy access to cross-references either within the article or to other articles on the disk or network. Quick, comprehensive cross-referencing clearly has advantages for a reader who might require a list of related entries (biographical information, reviews, etc.), or may wish to read and keep a printed record of the selected text. The SALit Web exploits the medium by accelerating the type of searching and cross-referencing already familiar to researchers and can even preserve a particular search "route" so that it might be used again later.

Despite the benefits offered by a single electronic text title, literary scholars demand a whole range of related texts at their fingertips, for quick referencing, cross-referencing and annotation. But at the moment most publishers do not offer the full scholarly potential of the medium because they tend to be unwilling to produce a text that implies sharing of information outside the proprietary "package" of the work itself. Any electronic publication may be a hypertext - with embedded links that the enable the user to navigate from one discrete chunk of information to another by activating highlighted text in the original document with a pointing device. The reader may just as readily use this feature to reach out via embedded links in the text to related print, archival, electronic and critical material located elsewhere. All such texts share a common digital medium, that may be searched for references, common words or word patterns, especially if they use a dedicated textual analysis application like the University of Toronto’s Text Analysis Computing Tools (TACT). Within the bounds of copyright, the SALit Web does make the texts in its Library section available for searching and analysis, and each text is hyperlinked to bibliographic entries (author, title, date) for easy identification. But embedded links could just as easily permit the reader to move beyond the perimeter of the SALit Web to remote text databases, archives, libraries and collections anywhere in the world. Hypertext coding (embedded links) within the full texts and in the bibliographic entries give the user access to a world-wide library of associated electronic texts. Although this potential versatility depends on textually-based hyperlinks and descriptive mark-up (document structure and word tagging, as in the TEI scheme), the tagging does not necessarily obstruct readability because electronic texts can be made to emulate the attractive format and page layout of their print-based counterparts: the Oxford English Dictionary Second Edition is one notable example of the happy co-existence of print and electronic versions. The user of the CD-ROM dictionary may wish to search for the earliest quoted word beginning with "technolog": in response to enquiry the dictionary opens a "window" displaying the appropriate page in the distinctive typeface of the Oxford University Press documenting the 1615 use of the word "technologie" by Buck in Third University English, p. xlviii (Seaman, 1994). Using the SALit Web, users would be able to access information like this from their desktops. As an electronic artefact, the SALit Web is a type of meta-text incorporating a hierarchy of linked material: primary entries including literary biographies, a chronological display of titles, historical periods, discursive and thematic formations, etc. that in turn contain cross-references, bibliographic references and full texts. A hypertext structure is non-sequential, and although the SALit Web material is hierarchically organised, progress through the material is user-driven.

Many scholarly texts, including encyclopaedias, already have a hypertextual structures where any table of contents, index, glossary, footnote or bibliographic reference implies reader movement (or "navigation") amongst several texts both within the printed document and outside it. The reader of a scholarly document may be prompted by a reference within the text to leave the main text to read a footnote providing additional commentary or information about sources, biographical or historical background. This, in turn, may lead the reader to another text (such as a critical article or a contemporary history) "outside" the original document itself. Conventionally, such a step would mean accessing the referenced document through a library. The reader might then return to the main text again. This type of reading may be characterised as "constitut[ing] a mental model of hypertext." (Landow & Delaney 1991: 5).

However, in a print-based environment (i.e. reading a book) even though navigation through associated texts is central to the reader's activity, the process can be protracted "because the references (or linked) materials lie spatially distant from the reference mark." (Landow & Delaney 1991: 5). If one were looking for the rare or marginalised South African texts included in the SALit Web Library, the difficulties of access to a printed copy could extinguish the process completely. The Oxford English Dictionary search illustrates the ease with which documents may be accessed in a hypertext network. Were the search to have been performed in a full-text database the scholar could have pinpointed and retrieved the full text of the novel, play, poem or article to study the context in which the word was used by all authors whose work is stored in the literary corpus. Wolff (1994) described how he used the ARTFL (Project for American and French Research on the Treasury of the French Language) database, to test his assumptions about the use of the words "real" and "ideal" in nineteenth century French prose:

In order to pursue the question of realism and idealism in the nineteenth century, one could begin to look for presuppositions about the about the real and ideal in texts from this period. [The database lists] words that often function to introduce presuppositions: quand, moment, quant, lorsque, etc. The keywords in this list occur in sentences that contain words matching the patterns. (Wolff, 1994).

If we now re-focus our view of the SALit Web, the overview diagram in could highlight the potential offered by the Library texts for textual analysis:

In Figure 3, a search for the word "river" (incidentally the highest frequency "place" designation word in Pringle’s Narrative of a Residence in South Africa") immediately displays the use of the word in several South African texts. In his account of using the ARTFL database, Wolff goes on to warn that it would be impossible to map any discursive space definitively, but electronic text resources allow the scholar to explore intertextuality in a manner impossible before the advent of computers. While the tools available for the analysis of digital text can be very helpful to the researcher, they can play only a mechanical part in the analytic process. By generating re-orderings of the original text according to semiotic categories determined by the researcher, linguistic and literary structures can be studied more quickly and more comprehensively, but the tasks of synthesis and interpretation cannot be delegated to the computer.

If the hierarchy depicted earlier in is extended to trace the associative direction suggested by, for example, "Bibliographic Entries", a reader either directly or indirectly liked via the computer to SABINET (South African Bibliographic Network) the diagram could be extended as follows:

In the example shown in Figure 4 , the reader is prompted by a pointer in CD-ROM to an electronic access route - SABINET and then the Online Computer Library Center (OCLC) - through to the designated article. But if the electronic (CD-ROM) version of the encyclopaedia were encoded for the purpose, it is entirely feasible that an appropriate hypertext link embedded in the bibliographic entry would perform the same activity automatically, following each of the steps in Figure 4 and culminating in the appearance of the full text of the referenced article directly onto the reader's computer screen. Internet users, for instance, already perform similar searches as a matter of course. In the case of the SALit Web, the reader could then read the text, and by using the built-in search facilities, either navigate through the retrieved text to the page included in the citation, or failing that perform a word, phrase or a "keyword-in-context" search (see Wolff, above) to target the desired content.

It must be stressed that the automated search activity described here is by no means whimsical or far-fetched. For the last five years undergraduate students at the University of Virginia have been able to query extensive text corpora in just this way from computers located in their residence dormitories (Seaman, 1994). The development of electronic text centres in the humanities has made available both the software and the texts to perform automated searches with flexibility and simplicity. The SALit Web strives to offer the same access to South African texts.

In the light of the research into electronic texts mentioned here, planning a resource such as the SALit Web without considering its extensibility into a network of associated electronic sources would be as bizarre as proposing that the work be hand-written on vellum. An examination of current trends in the development of electronic encyclopaedias suggests that the South African Literary Encyclopaedia, out of which the SALit Web project grew, is likely to become a CD-ROM publication itself. A digital version of the encyclopaedia would ensure that the scholarly activity invested in its production were optimised by utilising advanced computing tools. By taking the electronic route, the Encyclopaedia would avoid being doomed to an anachronistic mute existence, cocooned in an envelope of ink-text too costly for most South Africans to afford, or worse, locked into an incompatible publisher-specific encrypted code, capable of effecting only page-layout and formatting (procedural mark-up as opposed to descriptive mark-up).

How then would one position the encyclopaedia within the larger realm of electronic publishing - the "docuverse" (Landow and Delaney, 1991) or the "Information Omniverse" (Seaman, 1994), to use hypermedia jargon? The following diagram is an attempt to provide a perspective that defines the parameters of the specific SALit Web "package", while locating it within the wider network of existing (and projected) electronic text resources:

The above model allows the SALit Web as a distinct identity as a publication within the docuverse, while leaving links open to external associations. The hypertext possibilities differ only in terms of scale and physical constraints: the maximum capacity of one CD-ROM is approximately 600 megabytes, while remote e-text collections offer almost limitless storage capacity. Both intertextual and intratextual analysis of texts, as described earlier, are possible. A user may search for cross-referenced texts and record the results, or search for individual words using software - Folio Views or TACT - equipped with a KWIC "keyword-in-context" feature. Both the SALit Web package and the wider networks shown in allow for the close analysis of individual or large corpora of texts (e.g. TACT referred to above). One should imagine two distinct (though related) levels of activity: searching for relevant commentaries, critical articles and archive materials across a wide range of texts as part of a literature study, or the analysis of a specific text to test text-critical assumptions. These "levels" apply both within the SALit Web and the wider network, but the latter implies the existence of a body of full texts in electronic form that far exceeds the present collection in the SALit Web Library. Ideally, the SALit Web should be hyperlinked beyond its immediate domain to other collections of electronic text via the Internet. Even though from an electronic point of view, these would all fit seamlessly into the wider docuverse, for practical purposes they currently inhabit separate domains (the "gateway" between the SALit Web and SABINET, described above illustrates the point). For copyright reasons alone, it would be necessary at first to keep the SALit Web behind a password-protected gateway, just as the SABINET resources are.

We will narrow our focus to the SALit Web on CD-ROM and follow a limited number of possible search procedures, and in so doing show the current organisation of the interface.

Step 1: The Opening Menu

In Figure 2 (Overall structure of the SALit Web), the opening menu choices are represented schematically, with the highlighted arrows demonstrating one likely route through the Keyword section to the full texts. Immediately a major challenge confronting the web designer becomes apparent in the appearance and layout of this page. Several different opening pages were experimented with during the development of the web, including the map below, where a blend of geographical, chronological and historical icons present a very different conceptual impression of the web to the user than the one on the present CD-ROM. Clicking the image of the printing press would open a menu of references to the Lovedale Press and other links to South African printing and publishing, for example. The map metaphor has been used very successfully in the The Atlas of Literature (Ed. Bradbury 1996) and may yet appear as a navigation device as the web develops, but was considered too enigmatic for the students identified as the primary target group for this version of the SALit Web. The non-sequential nature of hypertext would just as easily lend itself to an opening page that depicted page one of the full text of a canonical South African novel, encouraging the user to read on, thus relegating the options of searching the database, etc. to a less prominent position on a sidebar. Similarly, even the overview diagram of the SALit Web (Figure 2) could be hyperlinked to the appropriate parts of the web, providing an opening menu that conveyed a much more compelling sense of the web-like structure of the resource.

The opening page finally chosen for the present CD-ROM version of the web (Figure 7) uses a textual menu more likely to be familiar to new users of the resource, and simply lists various user interfaces: the database, a full list of key words (search terms), the Encyclopaedia/ Dictionary, Library, Period Tours, Games and Backpack. To illustrate its use, let us search for bibliographic information on Thomas Pringle by choosing the "database" option. The composite "open" search form used in earlier versions of the database is not available in the current Beta 01 CD-ROM version of the SALit Web, where the interface has been simplified to automate some of the more likely search strategies of a student user, hence the appearance here of an "intermediate" level of search options:

Author search

Date search

Book title search

Article title search

Keyword searches

Whereas the original "open" search form required that the user learn several steps such as identifying the Author field, selecting it, and then using the MS Access search button to perform the query, the new level leads directly to the five pre-arranged menus listed above. As already mentioned, the advantage of the this design strategy is ease of use for the novice who is confronted with, for example, an alphabetical listing of author names, rather than a bewildering array of fields in which the position of the Author field may not immediately be apparent. Of course the disadvantage is the reduction in the number of formal field categories from which a more experienced user might make her selection at the start of the search process. Formal items that do not appear here include publication details, ISBN numbers, and gender but most of these fields become available once the user has reached the main bibliographic entry. The "open" search form will be re-introduced into later versions of the SALit Web. Clicking on the Author option opens up the display shown in Figure 8.

Once the user has found the author’s name, clicking the "click here" button opens the bibliographic form which contains the full list of the author’s works, conventional bibliographic details, a photograph of the author where available, and a short biographical sketch in the style of an encyclopaedia entry. Within the title listings are hypertext links to full texts in the Library section. Navigation between the bibliographic form and the full text requires only that the user click the Full Text hypertext "anchor" that in turn opens a new window containing the full text. By closing the full text window, the bibliographic display is restored. Returning to the main menu of the database can be accomplished either by selecting the "back" arrow icon in the body of the display or by closing the active window in the normal way. Bibliographic displays for any author together with all the functionality described above can be reached via any of the routes (date, book title, etc.) listed on the database main menu.

By far the most versatile of the choices offered on the database menu is the Keyword option. Our decision to index each bibliographic record with several different key literary concepts has elevated the query capabilities of the database to that of an encyclopaedia. Keyword coding of each record effectively brings to the SALit Web a meta-classification based on the combined erudition and scholarship of the research team members. The resulting displays therefore link often highly disparate authors, titles and dates in terms of identified themes and ideologies, as well as the more conventional categories like genre and language. Janus (1997: 607) calls this a "blend of tradition and new technology" in which the accumulated knowledge of the literary scholar cum indexer propels the formidable processing capacity of the computer. All Literature Webs share this general characteristic but it is the topic or keyword classification that exploits it most effectively. To take the idea further, the descriptive markup or encoding of full texts represents a similar combination of academic expertise and the computer’s ability to sort, list and analyse with great speed and efficiency. Within the SALit Web, there is categorisation at the macro level corresponding with the Keywords, and at the micro or document structure level, descriptive encoding of the full texts in the Library. TACT provides the means by which a user may "classify" or tag other structural features of the texts at a lower, more granular level. On the selection of the Keyword option in the database opening menu, the user of SALit Web is presented with a selection of topics, categorised in the current version under the headings of: Critical writing on ... (individual authors); Genres; Languages; Themes; and Ideological forms.

Some of the keyword topics are shown in Figure 9 (Click image to enlarge):

Inextricably coupled with the expert direction available to the user through the keyword features, is hypertext’s inherently democratic capacity to resist the primacy of any one voice in the text. The user is free at any time to ignore the pathways indicated by the keywords, and should she discover any new categories, mark and link these herself and so construct alternate conceptual associations in the network.

Two more of the Opening Menu choices are described here: the full text Library and the Period Tours. The other options on the menu are still in early stages of development and are not operational in the current version. The Library option offers the user access to the full texts without using the search features of the database and consists of an alphabetical listing of the titles and publication details. The benefit of the Library contents display is that it shows at a glance the current extent of the full text collection in SALit Web numbering some forty-five titles at the time of writing. The title in the list is linked to the full text, and clicking the title has the same effect as choosing the Full text links in the database bibliographic displays described above - the full text is opened in a new window. Even as the collection list grows, it remains a familiar and accessible way of perusing the available titles. Most Literature Webs and virtual libraries have similar displays which the user can scroll through, or perform simple searches with the "find" tool of a browser or other display application.

Among the texts currently accessible in the library are:

D.F. Bleek, The Mantis and his friends

W.H.I. Bleek, & L.C. Lloyd introduction to Specimens of Bushmen folklore

Roy Campbell's article "Fetish worship in South Africa: A skirmish on the borders of popular opinion"

Olfert Dapper's Early Cape Hottentots

H.I.E. Dhlomo’s "Literature and variety of tribal drama"

James J.R. Jolobe’s Poems of an African

L.C. Lloyd A short account of further Bushmen material collected

I. Nhlapo’s pamphlet Bantu Babel on the language question from Edward Roux's 1944 sixpenny library series.

Olive Schreiner’s Trooper Peter Halket of Mashonaland

While the Library option places the user and the texts together without any significant intervention from the web editor/designer (apart, of course, from the silent "canonisation" implied by the inclusion of those particular titles), the Period Tours are overtly expository in nature. Like a personalised rendering of the topical keywords feature, the tour guides the user through an illustrated account of the selected period, with hypertext links to associated full texts, reviews, photographs, maps, sounds, video clips and historical background relating to the time. Although the linear surface structure impels the reader from introduction to conclusion, the hypertext milieu offers multiple tangential explorations. Although the entire SALit Web is a hypertext, the Period Tours take the most advantage of the non-sequential essence of the medium, offering the user a truly open-ended opportunity to read, pause, examine related texts in depth, return to the Tour or simply skim through. The Period Tours available on the current version of the CD-ROM consist of historical surveys of five main periods of South African literary history:

0000-1652 (from Time Immemorial to the arrival of Van Riebeeck)

1652-1795 (from Van Riebeeck to the first British occupation)

1795-1900 (British Occupation to the beginning of the 20th century)

1900-1948 (from beginning of the 20th century to the institution of apartheid)

1948-present (Apartheid and post-apartheid)

Under each period tour menu item there is a link to the illustrated tour itself, to significant dates of the period, a dictionary of significant words related to the literature, an index of significant words and various important discursive formations. Currently the most developed Period Tour the first one in the list, covering the period from "Time Immemorial" to the first European colonisation.

Let us examine some of the features of this first tour "0000-1652 (from Time Immemorial to the arrival of Van Riebeeck)". The "tour" is a condensed narrative survey of the history of the period. Embedded in the tour text are hyperlinks to texts referred to, or to the topic under discussion, with important words and sections highlighted in red typeface. For example by clicking on "shamanism" next to Clashing of the rocks the user is taken to an excerpt from Chapter 9 of Lindsay’s study and clicking the "sound" icon, plays back a short recording of San speech. The text is further illustrated with pictures and segments of primary texts (indicated by a colour frame) and is sometimes accompanied by translations (see Marais’ "Die dans van die reën"). The immediate accessibility of translations is a major contribution made possible by the hypertext environment to a multi-lingual literary resource like the SALit Web and future revisions will incorporate English translations of African language texts into the library, so as to broaden research opportunities into this neglected area of South African literary study.

As can be seen in the above illustration, it is not only the hypertext features that may be exploited fully in the tours, but multimedia elements as well. During the development of the SALit Web, a small number of sound and video recordings were included for the purpose of experimenting with the full capabilities of the medium. The priority in the first phase of the CD-ROM development was the inclusion of a comprehensive top-structure of bibliographic references, while scanning and preparation of the Library full texts was second in importance. The creation of an image library ranked third in the list of priorities and now consists of over 500 pictures, mostly photographs of authors. With only primitive equipment at our disposal, the process of finding and converting sound and video recordings into digital form has been postponed to later in the SALit Web project development, although we have included the few sound and video images that were digitised on the CD-ROM. However, one of the more extended multimedia experiments deserves mention here. Designed as a spoken tour, "The Shamanic Formation in Southern Africa" consists of a five-minute illustrated commentary on the shamanic form, and in particular the trance dances of the /Xam. The tour was produced in MS Powerpoint format and was originally hyperlinked to the "0000-1652 (from Time Immemorial to the arrival of Van Riebeeck)" Period Tour. Using images of authors, rock paintings and other illustrations from the image library, together with a recorded commentary from a script written by Johan van Wyk, the Tour "talked" the user through the subject area in the equivalent of a short film or tape-slide sequence. Although it played as one seamless presentation, the tour consisted of 35 images (some of them animated) and eleven sound files. However, because the tour was likely to draw too heavily on the computer memory and processing resources of likely users, it was excluded from this version of the SALit Web.

Images and commentary from the first part of the presentation are included here by way of illustration:

(Van Wyk and Stewart. 1997)

"The Shamanic Formation in Southern African Literature" tour revealed the considerable scope that exists with in the SALit Web for the development of multimedia material. A solution to the digital format problems is available in the form of "streaming" applications like RealVideo, which we adopted for the sound clips on the CD-ROM. Only one session of original recording was conducted during the project: South African poet Dennis Brutus agreed to read a selection of his own poems and those of Arthur Nortjé for videotaping. One reading from this session "Silence in the still, warm room ..." was digitised and included on the CD-ROM. Additional sound and video clips include readings by other contemporary South African poets, and an extract from a filmed performance by a praise singer.

Earlier in this paper, we examined the relationship between the SALit Web and the rest of the "docuverse". As it stands, the CD-ROM version already has the capability of reaching out via hypertext links to the wider world of the Internet. But to be truly part of the "big picture", other users of the Internet need to be able to access our resource and this development marks the next phase of the SALit Web project.

Atlas of Literature, The. 1996. Ed. M. Bradbury. London: De Agostini Editions.

Dowining, D.A. et al. 1997. Dictionary of computer and internet terms. New York: Barron's.

Janus, R.J. 1997. "From paper and ink to CD-ROM: Digitizing the World Book image". Library Trends. 45:4. 602-622.

Landow, G.P. and P. Delany. 1991. "Hypertext, hypermedia and literary studies: the state of the art". Hypermedia and Literary Studies. Eds. G.P. Landow and P. Delaney. Cambridge, Massachusetts: The MIT Press.

Seaman, D.M. 1994. "‘A library and apparatus of every kind’: The Electronic Text Center at the University of Virginia". Information Technology and Libraries. March, 1994: 15-19.

Sperberg- McQueen, C.M. 1995. "Bare Bones TEI: A very small subset of the TEI Encoding Scheme". Document No. TEI U6. Chicago: Unversity of Illinois at Chicago.

Wolff, M. 1994. "Poststructuralism and the ARTFL database: some theoretical considerations". Information Technology and Libraries. March, 1994: 35-42.