V. Creating the Hypertext Cluster


When I arrived at the University of Kent at Canterbury (UKC) in 1992, I discovered that the University placed a very high priority on information technology (IT); in addition to a dedicated faculty, there was a computer centre which offered all students email accounts and access to personal computers. New students were automatically sent introductory packs, and this meant that I quickly came into contact with faculty members, including Wilma Strang from the Hypertext Support Unit. One of only two similar units in the country (the other being at Oxford University), the Hypertext Support Unit was founded to introduce hypermedia to the academic community and to develop hypermedia teaching materials. With Wilma Strang's assistance, I began to read theoretical work on hypermedia, most notably by Professor George Landow of Brown University, Rhode Island (whose work I thoroughly recommend to other researchers). I also benefited from the work being carried out at UKC: for example, one of the first and most accessible academic hypertext systems, GUIDE, was developed at UKC by Dr Peter Brown. By 1993, I had therefore become familiar with the concepts of hypermedia and electronic publishing. (A full introduction to and discussion of this can be found in The E-Book & the Future of Reading.)
ptr1

I had written my MA thesis using an Amstrad wordprocessor, so the final, bound version had essentially been a computer print-out. In order to meet the criteria for examination, I had printed the thesis on A4 paper using double-spaced text with wide margins, and the Amstrad's dot-matrix printer meant that the type was grey and not clearly defined. As a journalist, I knew that visually the MA thesis was much harder to read than a book or magazine. Equally, as a professional researcher and writer, and having already produced a traditional 50,000 word thesis, I did not feel that producing another traditional thesis provided much of a challenge for three or four years of study. I knew that, whatever I did, I would write my thesis using a computer. What would happen, then, if I designed it to be read on a computer? I was already intending to record my oral research on video, and I knew that I would have a large number of images to include as appendices. On the computer, I could integrate all of these with the text, creating a hypermedia thesis or hyperthesis.
ptr2

Hypermedia is characterised by its non-linear, open-ended structure, and conversely by its linked nature. Presenting my research electronically would therefore allow for a greater breadth of research to be presented than in a traditional linear thesis. Feminist research is characterised by crossing borders between traditional academic disciplines, and Communication & Image Studies is in itself a multi-disciplinary area of study, so this seemed to be particularly appropriate to my research field. The nature of my research, too, had already made it inevitable that traditional academic boundaries would be broken. And as the subject of my research was books and reading, and I was looking in detail at the part which books played in girls' and women's lives during the twentieth century, it seemed highly appropriate that the form and presentation of my work should explore the future of books and reading in the twenty-first century.
ptr3

In developing a hyperthesis, I could also realise the theory which had been developed around hypermedia within my academic field of Communication & Image Studies. As Landow points out: "hypertext has the potential to serve as a laboratory for theory while theory illuminates the design, use and cultural effects of the new electronic technologies." At the time, the majority of non-fiction hypertexts which were available were based on the encyclopaedia or "database" model, while the leading-edge work being carried out at Brown University and others was based on the "web" model. In both these cases, the concerns and interests of the reader were paramount in deciding reading paths, reflecting theories about reader-empowerment and non-linear narratives. In contrast, I felt that it was possible to combine the best of both linear and non-linear structures.
ptr4

A hypermedia thesis would not, of course, have the single controlling line of argument which typifies the traditional, printed thesis, but I did feel it was possible to keep sight of the fact that my research was aimed at finding a fundamental schema of analysis. My work would therefore represent the first attempt at creating the electronic social sciences/humanities "textbook". The hyperthesis would then demonstrate hypermedia to readers who had not encountered it before, and within it I would explain the design process so that other researchers would be assisted in producing new hypermedia texts.
ptr5

Since the form would be new to most "readers", I decided to include a separate, but linked, hypertext, which discussed the impact of hypermedia on the book and the future of reading. While this writing would be appropriate to my academic discipline of Communication & Image Studies, it would also have a direct connection with my research into the part which books play in their readers' lives. In general, the decision to present my thesis as a hyperthesis allowed me to present other texts alongside the core thesis on girls' school stories, underlining the fact that a hypertext is a collection of related texts rather than a single narrative. Whereas in a traditional thesis these other texts might have been given the status of appendices, in the hyperthesis they "gained an importance" which they would not have had before, since "in hypertext, the main text is that which one is presently reading". However, since the "appendices" all had separate narrative focuses to the main thesis, I decided that it was more appropriate to describe the work as a "hypertext cluster" rather than a single hypertext. While the "main" thesis represented the culmination of my research findings, the "appendices" represented work which I had carried out during the period of my research.
ptr6

One important element which the structure of the hyperthesis allowed me to include was autobiographical. First, I could introduce my personal history of reading within the Foreword. Reinharz points out that:

ptr7

I also decided to include a collection of texts within an autobiographical lexia, About the Author. Liz Stanley and Sue Wise define two principles of feminist research as being situated "in emotion as a research experience" and "in the intellectual autobiography of researchers". They add that:

ptr8

Similarly, Reinharz points out that feminist research is characterised by the belief that it is important to provide information about the background and experiences of the researcher. "Some feminist social researchers have written full autobiographies, or have written full reports about their experiences as researchers of women . . . more commonly the researcher adds a preface or postscript that contains an explanation of her relation to the subject matter at hand." Including an autobiographical lexia would also encourage readers to speculate on the place of their own autobiography in structuring their responses to popular culture.
ptr9

I decided to provide brief details about my professional background in the shape of a CV. I also decided to include a much longer piece of writing about my own experiences at a girls' selective state school in the 1970s, which became the linked hypertext My Own Schooldays. I wrote this partly in order to provide further autobiographical evidence, and partly because no published accounts existed of similar experiences, and I believed that it would be important to consider the differences between real and fictional school experiences elsewhere in the hyperthesis. Maggie Humm points out that: "One hallmark of contemporary feminist research in any field is the investigator's continual testing of the plausibility of the work against her own experience", while Reinharz states that:

ptr10

Later, I also decided to include a separate, but linked hypertext which discussed my experiences as a disabled researcher. Although this was not directly relevant to my research topic, it was directly relevant to my experience of carrying out the research, and therefore as a feminist researcher I felt that it was important to include it. Reinharz points out that: "Many feminist ethnographers have eliminated the distinction between the researcher and the researched and have studied their own experience.". She adds that:

ptr11

Another reason for including my experiences as a disabled researcher was the fact that it allowed me to address ongoing debates about the body within my academic field of Communication & Image Studies which were not directly relevant to girls' school stories. And the fact that disability has had such a low profile within the academy in the past provided further justification from a feminist research perspective. As Reinharz points out:

While Reinharz is not discussing disability here, and her words are highly relevant to the subject of my research, girls' school stories, they are equally relevant when arguing for the visibility of a disability perspective on the research process itself. (Earlier, Reinharz points out that "feminist content analysis is a study both of texts that exist and texts that do not".)
ptr12

(Discussing my personal experience of the research was an approach which I would also adopt throughout the hyperthesis where relevant. Reinharz points out that:

Humm adds that: "women researchers have to begin with personal experience . . . For women an emotional reaction is . . . often the foundation of critical thought and more astute theory".)
ptr13

I also decided to include part of the text of my MA thesis within the hypertext cluster, since the MA represented the first stage of my research and only five bound copies exist. (Initially there were four, but I had an additional copy produced when it became obvious that many collectors wished to borrow and read it.)
ptr14

As the presentation of my findings in a hyperthesis marked a new stage in the development of the book and in the presentation of academic research, and as research had not been previously carried out into reading experiences associated with girls' school stories, I had to develop new ways of working and to use new research techniques. This included the use of digital video editing, desk-top publishing and multimedia authoring software. After completing this research, I decided to include a separate, but linked hypertext containing the information which I had discovered about health and safety practices when working with new technology. I decided to include it first because of its relevance to the practice of reading Virtual Worlds of Girls, and second because of the widespread ignorance which currently exists about the safe use of computers. This hypertext has already been published in print form by Skillset, the training body for the film and television industry, as Health and Safety in the Non-Linear Environment (1995).
ptr15

Other linked texts within the hypertext cluster include the "notes" to the "main" texts. While some are simply references or short quotes or notes, as with traditional foot- or endnotes, others include extended quotes or detailed discussions of lateral issues. Landow points out that, in printed books:

ptr16

One aspect of research which hypermedia makes explicit is the fact that no piece of research exists in a vacuum; it is all inter-connected. In an electronic medium, these connections can potentially be made explicit by creating links between texts. Many of the "notes" within the hyperthesis are in fact potential links, whether they simply supply a reference or include an extended quote. Conversely, when texts can be linked, there is no need to summarise another piece of research unless it is directly relevant to the topic under discussion.
ptr17

Ultimately, I also wanted to include the presentation of my video material, The Chalet School Revisited, as a separate lexia, having discussed the making of it within Exploring the World of Girls' School Stories. In addition to this being an integral part of the hyperthesis, when viewed on computer, readers/viewers could selectively view individual scenes or tracks and decide the order in which they were played.
ptr18

I could, of course, have also included my research notes and records within the hyperthesis, but I decided that this was inappropriate. However, I did resolve to archive all of the material, and in the future to make it available to scholars on request where possible. I also finally decided against including sections which merely provided a historical record - for example, excerpts from fanzines; details of ephemera which had been produced by the fans; and fanclub histories - and which should in any case be kept separate from interpretation of these phenomena. This was not because these records were irrelevant, but because by 1996 the fans themselves were now capable of producing and publishing their own records, both in print and electronically.
ptr19

Developing the hypertext cluster meant that I was now effectively presenting two theses: one which presented research on girls' school stories and their readers, including the memoir of my own schooldays which could be compared to the educational experiences represented within the genre; and one which was concerned with the research process itself, including Exploring the World of Girls School Stories, The Ebook & the Future of Reading, autobiographical details, My Experiences as a Disabled Researcher, and Health & Safety in the Non-Linear Environment. In the first thesis, my research into the genre of girls' school stories was the subject; in the second, my research was the example which illustrated the process itself, including the development of a new means of presenting research.
ptr20

Structurally, I found that there were two challenges when writing the hyperthesis. First, while it would be quite possible to divide the text into virtual "pages" which could be turned, there was no point, since it would hamper the reading process, and in any case, pages generally create artificial divisions in the text which were not intended by the author (an exception to this are spreads of illustrated pages). I also considered it better to create one larger file size which contained a chapter or lexia than a group of smaller ones which each contained a page, given that on the World Wide Web the reader is forced to wait each time they access a new file. However, the use of paragraphs marks divisions in the text which are intended by the author, so explicitly dividing the text into paragraphs rather than pages seemed to be a more "writerly" way of doing things. In order for the reader to be able to reference the work, I included a reference number at the end of each paragraph to replace the traditional page-numbering system.
ptr21

The second challenge which I encountered when writing up the research was that the ability to break down text into small self-contained chunks meant the text as a whole lacked coherence at first draft stage. At second draft stage, I therefore reassembled many of these chunks into longer, more linear pieces of writing. I am, however, convinced that with a complex piece of work such as this, with non-linear reading paths, the author does benefit from going through this process and examining the many different relationships which exist between chunks, before deciding how to structure the final work.
ptr22

Equally, at the time of writing (1997), most readers had only encountered hypermedia in the form of World Wide Web pages where text chunks are designed to be as brief as possible, so it was important to consider how best to reconcile the expectations of experienced Web users and those of print readers. At the same time it was important to bear in mind that the Web is an experimental means of communication which had far exceeded the expectations of its designers, and to see current developments in hypermedia as representing a transitory rather than final stage in the (multi) medium. I therefore gave greater weight to theoretical considerations than to contemporary examples when designing the final structure.
ptr23

At the point when I developed the idea of creating the hyperthesis (1993), I decided that I would not concern myself too greatly with how I would produce or "author" it. If I made initial choices about which software I would use, my design would be constrained by its limitations. It would be better to decide on the design now, and then to choose the most suitable software available when I came to author it, modifying the design at this point to take account of the chosen software's characteristics and limitations. In any case, with new software constantly being developed, it was likely that the software which I would eventually choose had not at this point been released to the public. Landow later described this to me as "designing ahead of the curve", which he personally recommends.
ptr24

At this point, too, the World Wide Web was in the process of development, and writers were using a variety of programmes, such as Guide and StorySpace, to author their hypertexts. These programmes produced hypertexts which were, on the whole, inaccessible to people who were not already familiar with computing conventions, and which could only be used by people using the same software and system as the writer. And while copies of these hypertexts could often be obtained via the Internet, they could not be read online. I therefore conceived the hyperthesis originally as a standalone programme - one which did not require the authoring software to read it - and with a custom-designed interface.
ptr25

At the time when I began writing up my research, I had already created the Bettany Press publications in QuarkXPress 3.31. I was also aware that a multimedia XTension to the forthcoming QuarkXPress 3.32, Orion (since retitled Immedia), was in development. Corporate literature distributed at the 1995 Apple Expo (held in December at the Olympia exhibition centre in West London) promised that Immedia would combine sound, still images and moving images with text, be "easy to use", be a "multiplatform solution" (since the products created could be viewed either on an Apple Mac or a IBM PC running Windows) and that products created with it could be viewed over the Internet. I therefore decided to use QuarkXPress 3.32 to write up my research, since the documents could always be translated into another format if I changed my plans.
ptr26

In the event, by 1996 the success of the World Wide Web meant that my aims would be better fulfilled by presenting the hyperthesis as a web site, which could be read in whatever browser (Netscape, Internet Explorer etc) the reader was familiar with. I therefore translated the documents back into Microsoft Word and then saved them in RTF (Rich Text Format) before translating them into HTML using the shareware programme TextToHTML 1.3.4. I then began to author the hyperthesis using Netscape Navigator Gold as the main editing programme, with Microsoft Word for editing the code where necessary.
ptr27

Bob Cotton and Richard Oliver set out four main principles for creating hypermedia products:

Taking the last point first, it is clearly important to have a working knowledge of the possibilities offered by the software and hardware being used, and to have enough knowledge of the different types of software and hardware available to have made the best possible choice before beginning. However, it is not necessary to be an expert: I was able to use Quark XPress to create the first Bettany Press book with only a few hours of training and to edit The Chalet School Revisited after a three-day course; in both cases I was new to the Apple Macintosh hardware which I was using and had not previously used editing or desktop publishing software. As George Landow recommended to me, the best way to gain knowledge is "learning by doing". And it is extremely important to create the initial design without reference to the limitations of the available software - Landow's "designing ahead of the curve".
ptr28

Returning to Cotton and Oliver's first point, the utilisation of different media to do what it does best, I wanted to use images to "show" rather than simply to "tell" the reader where this was relevant. This seemed to me to be more in keeping with the empowering and open nature of hypermedia than using a description which the reader has no choice but to accept; it also seemed more in keeping with feminist research methodology. I had of course, used video to record the Elinor M. Brent-Dyer centenary events, and had edited this into the 60-minute film The Chalet School Revisited (I discuss the presentation of the video material elsewhere). I had structured the film so that, like the printed text, it could be broken down into chunks, in this case scenes. When viewed on computer, readers/viewers could selectively view individual scenes or tracks, and decide the order in which they were played. However, in 1997 the majority of computer users did not have access to the necessary equipment to play full-screen video, so for the time being I had to present the film separately on tape, with links within the hyperthesis made instead to the script.
ptr29

However, I was aware that the facility to compress and store large files of moving images would soon be available, and that this would raise ethical as well as structural questions for future researchers. Theoretically, with enough storage space, I could have made my original footage (the "raw" footage) available alongside my own edited presentation of the material. Readers/viewers could then have chosen what to view and in what order, and to see footage which I had excluded during the editing process. They could also - if I desired - have the ability to edit their own presentations of the material. This would have shifted the balance of power almost completely from myself as an author/editor to the reader/viewer/editor, although I would still, as the photographer, have determined what material existed in the first place. But while the described potential loss of power from the author/editor to the reader/viewer/editor is to be welcomed, what effect would this have on the research subjects?
ptr30

To some extent, those being filmed for my research were aware of the final form in which this footage would result; they were aware of my motives for filming them and agreed to be filmed on this basis. Although they had little power over the final result, they could make their initial decision to participate by making judgements about the author/editor/researcher and could continue to participate on this basis. But if I included the raw footage, the subjects would lose even more power over the destination of their images. I would therefore have decided against including it even if it had been technically possible. Nonetheless, I could use still images from the video material to illustrate the text, along with relevant photographs. Throughout my research, I had taken photographs using a 35mm SLR camera, in this case a 1987 Nikon F-401, generally shooting on 400 ASA transparency film.
ptr31

(In fact, a compact 35mm camera would have sufficed in most situations. I would not, however, recommend using a digital camera at the time of writing, although transferring images to a digital medium is obviously a much simpler process than with conventional film. But while digital cameras are extremely light and easy-to-use, the majority can store less than a hundred images at a time, leading to over-selection at the point of photographing. In addition, the image quality is often only suitable for viewing onscreen, and even when this is the primary purpose of the pictures, it may well be that further, print uses are also found for them.)
ptr32

Cotton and Oliver point out that:

Where possible, I used the Kodak Photo CD service to transfer photographs to CD Rom, before using Photoshop 3.0 software to crop and otherwise manipulate them (as Photo CD images are designed to be viewed on a television screen, they tend to be over-saturated with colour) before saving them as jpeg files. Since I also used photographs taken by fans, for example of my filming them, I tried where possible to obtain negatives or transparencies and to digitise them in the same way. However, if only prints were available, for example of craft work created by Australian fans, I used a flatbed scanner and Photoshop 3.0 software to digitise them directly into my own computer. I also used a scanner to digitise line drawings and other illustrations from the books, using the freeware programme Transparency to make the backgrounds of some transparent before saving them as gif files, since I wanted the reader to be constantly aware of the images of the genre. In terms of capturing still images from video, I was able to connect my domestic video player to the back of my computer and then use Photoshop 3.0 to manipulate the size and quality of the pictures before saving them as jpeg files.
ptr33

In terms of colour, I decided to retain the traditional black of printed text for the body of the text, partly for ease of reading, and partly to continue the association with print book design. For the same reason, I decided to keep the background to the text or "page" white, although I selected a "chalk" pattern to give it more texture as well as creating an association with the traditional school blackboard. I then decided to use blue (1919FF) for the headings and sub-headings, and red (FF1C1E) for the paragraph numbers. These were colours which were already associated with the research, as I had designed the associated Bettany Press print publications using red for the background of the cover of The Chalet School Revisited and gentian blue for Visitors for the Chalet School. I had chosen these because red and blue were used on the "boards" of pre-War first editions of the early Chalet School books, while gentian blue was also the colour of the uniforms in the "Swiss" part of the series. Blue was also reminiscent of school ink, and red of the ink which teachers use to mark work.
ptr34

Type sizes are only relative in HTML, so I used -2 for the paragraph numbers, +1 for the body type, +2 for the sub-headings and +3 for section headings. However, I decided to create the main headings as image files, in order to be able to select the typeface which was used for them. Bob Cotton and Richard Oliver point out that:

In this instance, I decided to avoid three-dimensional and animated typography, since the subject of the hyperthesis was printed books which are characterised by their two-dimensional, static typography. Instead, I chose CartoonArt for the typeface foor the main headings, since it is reminiscent of writing on a school blackboard: I used red for the titles of the core hyperthesis, and blue for the linked hypertexts. However, I decided against controlling the main type more closely: for example by specifying the width of the text; by controlling the amount of space ("leading") between the lines of type; and/or by designing for a browser which could specify the font in which it was viewed. In print, the designer must make these choices; the reader has no control at all. But in hypermedia, the reader can decide what is the optimum font and width of text for their own viewing, and I know from my own research that this varies widely. Taking these choices away is to betray a misunderstanding of the nature of hypermedia.
ptr35

Clearly it was important that readers could easily understand how to read the hyperthesis, and this meant that the interface design was crucial. Cotton and Oliver explain interfacing as "being able to control machines by communicating with them, and receiving feedback from them. Watching the speedo on a car dashboard and easing up on the throttle, or choosing the correct washing machine programme, are good examples". They go on to point out that:

Now that I had decided to present the hyperthesis on the World Wide Web, there was not a great deal which I could do to influence the Graphical User Interface, or GUI, since the bulk of it would be represented by the browser. I therefore decided to make the most of it by letting the reader use the "Back" feature on their browser to retrace their steps after following links, rather than putting this feature into the code.
ptr36

In terms of structure, Cotton and Oliver warn that:

And Landow points out that, as a result, the hypertext systems Storyspace and Intermedia "both provide forms of dynamically created concept maps. Intermedia automatically generates the Web View, a dynamically graphic concept map that informs the reader by means of labelled icons which documents 'surround' the document one is currently reading."
ptr37

However, Cotton and Oliver also point out that:

Since my original aim was to attract "those readers, especially women, who are not computer literate, and who have no professional or academic computing support", and my belief was that readers of traditional printed books are far more likely to understand book metaphors when using hypermedia than those related to "landscape gardening, sculpture or architecture", or indeed maps, I decided to use a traditional print-book, linear "contents" or index system. I also believed that book metaphors would be more appropriate to the research topic of girls' school stories. And, as Stuart Moulthrop points out in his "Axiom of Familiarity": "The screen is but a flickering page. While we work in words, we are still people of the book. Gutenberg lives."
ptr38

I decided against including the index itself permanently onscreen in a "frame", though, as so many Web users seem to dislike this design feature, and also against using "buttons" as links, since these are slow to load and add nothing to the reader's understanding. Instead I decided to include links back to the sub- and main indexes at the end of each lexia after a link to the next lexia in the recommended reading path. In terms of the appearance of the links, I decided against using the traditional underlined blue "hotwords", partly because they are visually so distracting, and partly because I wanted the reader to be able to distinguish between different types of links when deciding whether or not to follow them. I therefore chose to use images instead, in the form of differently coloured and directional arrows. I decided not to include textual alternatives to these, thus preventing readers from using text-only browsers, because I felt that the design was integral to the presentation. I would, however, include this feature in an adapted version for disabled readers, which would also be suitable for a touch-sensitive screen.
ptr39

I decided to use a black arrowTHIS IS A LINK TO ANOTHER PART OF THE SAME CHAPTER pointing right (in design terms, this means "next" in cultures which read from left to right) to indicate a link to another part of the same lexia; a red arrowTHIS IS A LINK TO ANOTHER PART OF THE VIRTUAL WORLDS OF GIRLS WEBSITE pointing right to indicate a link to another lexia in the Virtual Worlds of Girls hypertext cluster; and a blue arrowTHIS IS A LINK TO A REFERENCE FOR A QUOTE pointing right to indicate a link to a reference, used smaller than the actual size to underline the fact that in a true hypertext, the reader would be able to move instead to the actual text. I then used a purple arrowTHIS IS A LINK TO A NOTE PROVIDING FURTHER INFORMATION pointing downwards (to show that another "level" of information was available "below" the lexia being read) to indicate that a linked note/further information on the topic was available, with a green arrow pointing downwards to indicate that a linked picture was available. I chose these colours because they were the Suffragette colours, as well as having queer connotations, and are also two of my personal favourites.
ptr40


Next: My Experiences as a Disabled Researcher
Return to: Researching & Creating Virtual Worlds of Girls Index
Return to: Virtual Worlds of Girls Index

Dr Ju Gosling aka ju90's ABNORMAL: How Britain became body dysphoric and the key to a cure is available now for just 3.09 for the Kindle or in a limited-edition hardback with full-colour art plates for 20 inc UK postage and packing. Book cover