| CHWP A.6 | Best, "Dehumanizing the Internet: The Control of Display" |
A fundamental difference between the electronic text and the printed text is that the electronic text is dynamic, not only in the way the content can be changed at will by anyone with access to the digital version of the text, but in the way the electronic text is displayed visually. An obvious example is the way all word processors now provide the capability for changing font faces and sizes, as we are discovering from our students, who change the font to fit the number of pages required for an assignment. What is less obvious is the way that our word processors and other computer programs have indirectly imposed the limitations of the earlier medium of print on the new medium of electronic text: what we see on the screen is specifically geared to print output, and the more sophisticated the program, the more closely the display on the monitor approximates the way the document will appear when printed. This is the much vaunted feature we know as "WYSYWIG" (What You See Is What You Get). I have elsewhere pointed out that "what you get" is what you get on a printout, with the result that the image on the screen is every bit as limited to the medium of paper as a typewriter (Best 1995).
Since the purpose of most of our word processing is still to produce paper "hard copy," this is clearly an advantage, and those of us who have been using the computer as typewriter long enough to remember the arcane "dot" commands of early word processors are enormously relieved to be rid of them. But the material we view on the Internet is not usually designed for printing -- if users want to refer to the page again they will bookmark it or download it. The paradox is that the design of Web pages is increasingly being anchored to visual principles that are derived from print technology, and in the process they are making the Internet less flexible, less creative, and less responsive to individual needs or preferences.
Technology has always had the potential to reduce the personal and the varied texture of humanity. Nostalgically, we may regret the passing of the creations of the human hand as individually crafted artifacts stamped with the fingerprints of imperfection are replaced by machine-made, uniform objects; wormy organic food is replaced by the crisp product of a monoculture; handwriting is replaced first by the typewriter, then the wordprocessor, where we may suspect that our friends have sent us a form letter using mail merge to "personalize" it. It is hardly surprising that the technological has come to be associated with the impersonal, the less-than-human.
The electronic screen can offer a significant counterbalance to this trend in that it offers an opportunity for the recipient to become more actively involved in the process of reading than is possible from a fixed page. Not only can receivers control content through hypertext links, they can choose the visual format of display in their browsers, specifying default fonts, font sizes, colours and so on, suiting the display to their particular monitor, eyesight, or esthetic sense. Thus the potential is there for the expression of personality in a different way, as the receiver of the information rather than the sender/designer.
But does anyone actually customize the browser in this way? I have watched many of my students working on the Internet, and have seldom seen them even change the size of the browser window, let alone change the font. There is an interestingly contrasting decision by the designers of the major Internet browsers: Netscape Navigator, which is the most popular, requires that the user go deep into the menus to find "Preferences" and make changes to the default font and size in a complicated dialogue box, while Microsoft's Internet Explorer provides convenient buttons on the main toolbar for changing the font size of the display, if not the actual font face. This is not a trivial difference in emphasis. As soon as you change the font size or font face, you change the look of the page, and the designer of that page no longer knows exactly what it will look like. The difference between these two browsers is a reminder of the way that basic decisions of interface design can reflect ideology: Netscape provides several buttons on its basic toolbar that take the user directly to their site, but none to change the look of the screen.
Designers, educated on the printed page, want certainty; they want us to look at a screen exactly as they have planned it. [5] An Internet site at Microsoft discusses some of these issues in the process of introducing its range of standard fonts, using a new form of font description that will ensure the same typeface across different computer systems:
The Microsoft site continues, pointing out a further disadvantage of purely display-oriented sites:
The Microsoft site illustrates neatly another habit of designers determined to decide how their text will be viewed. The screen is divided vertically, by tables rather than frames on this occasion, and the width of the text is designed exactly to add up to the 640 pixels on the standard computer screen. If I choose to be cranky, and want to keep the window of my browser small so that I can work on more than one task at a time, I am forced into what must surely be the most inelegant and inefficient way of viewing text: the horizontal scroll, which turns the experience of reading into something close to watching a tennis match with frustration replacing excitement. [10] The assumption that users will use their whole screen is so common now as to be the norm.
So what? There is nothing obviously limiting in the choice of a representational tag over a logical one, but I would argue that there are two submerged choices at work, both of which are in a sense ideological: the imposition of patterns of thought from the older print medium, as I have already demonstrated, and, more disconcertingly, the choice by the providers of data to control rigorously the precise format of the display -- a familiar desire on the part of the owner to exercise power over the material provided. The newest versions of HTML, both official and unofficial, are providing codes for the definition of the width of the window and the graphics, tables, or frames within it in absolute terms by pixel. [Hypernote 2] And as soon as the code is provided the designers will use it. At our end of the data stream choice melts into air, and our dreams of expressing our own personality over the appearance of our screens evaporate. In a kind of paternalistic or corporate mentality in which the designer knows best, we end up not with WYSYWIG but with WYSIWWWYTS: What You See Is What We Want You To See. As users (or "consumers" as the rhetoric would have it) we send a small signal upstream -- a click of a mouse button over a link or part of a graphic -- and the designers send us a huge stream of data downstream as they provide us with a glossy image, complete with three-dimensional graphic effects, headings with obligatory drop shadows, animation, and a sound bite or two.
Even when site designers try to do something new, they seem destined to perpetuate some of the same rigidities of design. The home page for Coke tries to break out of the current clichés. It is a single image: on the left is an antique Coke urn, on the right a message scrawled (in the same bitmap image), wittily trumpeting the timeless universality of Coke compared with the postmodern contingency of Fanta, which changes flavour with culture. [11] If you realize that you are supposed to click on the urn, you will be taken to a page where the designer has once again decided that the only way to view is with a full screen; the margins are elegantly wide (forced by a triple nesting of the <blockquote> tag), and the effect is delightful -- if you choose to display it the way you are supposed to.
Visually, web pages like this make it seem almost as if we are sitting in our armchairs with a remote control, clicking at the screen to change the channel.
And that is no joke.
WebTV is a revolutionary new way to access the Internet from your TV. You don't need a computer and there's no software to load. All you need is a television, a phone line, and a WebTV Internet terminal, and you're on the Internet. [12]
If the older technology of the book is cramping the style of the new medium by attempting to make it behave like a printed book, the influence of another medium, television, may be about to have a similarly distracting effect on both style and content of Internet sites. If, as Michael Joyce claims, "hypertext is the word's revenge on TV", WebTV may be the counterattack of TV on hypertext (Joyce 1995: 47). Users who have a reasonably up-to-date browser will have found that many sites already insist on making things move as they try to read, thanks to animated GIFs and the use of JavaScript. No doubt the result of student days in the Southern Hemisphere preparing for exams on a beach sticky with heat, my own response to something moving on the surface of a text is to feel that I should squash it before I read. It is a curious reflection on the difference between image and text that the noxious <BLINK> tag has tended to fall into welcome disuse, while animation is spreading like an e-borne virus.
The mission statement for WebTV is laudable: "To make the internet as accessible and compelling for consumers as broadcast television is today."
The designers of WebTV are astute, however, in analyzing the current trends in Web design. They comment: "Up to now, most content developers have been designing pages very similar to a hard copy book or magazine" -- precisely the point I have been making. They also provide some acute comments on the likely preferences of an audience accustomed to the TV rather than the computer screen:
Reduce the number of items on your page -- television audiences are used to looking at one focal point. Next time you watch any television show, notice that your eyes are always directed to one particular spot on the screen. Although your page won't have just one element that directs focus, you can design your page with fewer items and with the most important item so placed on the page as to draw the viewer's eyes to it.
I do not believe, however, that these developments are entirely negative. There is room on the Internet for the glossy and impersonal as well as more content-oriented and flexible sites. If WebTV does win an audience, there is no doubt that some of those who access the Internet via their TV will find more than sound and word bytes as they explore. Indeed, the advent of the cramped, low-resolution screen may cause some current designers to be less rigid in their use of font faces, fixed screen widths, and so on, since the TV browser will not be able to scroll horizontally at all. There is no reason why the Internet should not be home to as many different levels of visual presentation as it already is to other kinds of discourse.
One of the greatest strengths of the Internet as a medium is that, as bandwidth increases, it is increasingly capable of absorbing all the older media -- print, the visual arts, music, television. But instead of using the multimedia capabilities of the Internet to create forms that more fully represent the complex content of modern culture, the stress has so far been on eye-candy. The gloss-to-signal ratio is rising as the new browsers encourage more rigid structures, whether they are derived from the printed page or the TV screen. There is a strong argument that the electronic medium requires fundamentally different design principles from those of the more traditional media. Not only is the Internet capable of all the advantages and challenges in the effective use of multimedia, but Net users read/look in a different way from the way they do as they look at a page or television screen: the revealing metaphor of "surfing" reminds us that the Internet user is a "hit and run" viewer, with an attention span often more transient than even the similarly labeled channel surfers of television. The challenge of the designer within the medium is to create new ways of orchestrating the dance between form and content such that the eye is offered more nutrition than candy.
There are some academic sites in the Humanities that are (in my view) falling into the trap of putting gloss before substance, and reducing the choice of the end users. Part of the problem of course is that humanists are not programmers, and HTML has become sufficiently complex that we must either rely on (that word again) WYSYWIG editing programs for HTML that decide for us the nature of the underlying code, or we get funding for programmers -- who love to program, and who therefore create wonderful, technologically ornate sites where the technology, far from enriching meaning, actively works against it.
An earlier version of the US Shakespeare Globe site at the University of Illinois was a good example of form triumphing over content (it has since been elegantly streamlined). The opening screen was very charming: Shakespeare was shown initially in a portrait derived from the familiar Droeshout illustration from the first Folio; then his eyes moved, his eyebrows were lifted, and he started to have a thought balloon, which turned by stages into the Globe Theatre in its 1997 manifestation (minus the conglomeration of modern buildings that in fact crowd it on every side). This was surely a case where technology was not in service of content, since there is no evidence at all that the Globe or its stage was the result of Shakespeare's personal inspiration -- it was almost certainly the thought balloon of the Burbages. [14]
The irony is that browsers are becoming more sophisticated in passing the innocently nicknamed "cookies" of information back and forth [15] so that the designer of a site can learn precisely what kind of system a visitor is using to view data; there is thus no reason why advanced sites should not provide multiple ways of viewing the text -- more than the perfunctory offer of a "text only" alternative. In addition, the capability already exists for varying texts to be generated "on the fly" by the site, according to the expressed preferences of the user. In the context of my own work on the Internet Shakespeare Editions, it is my aim to make it possible for users to select varying levels of annotation when viewing a text: unannotated, lightly annotated, or full scholarly annotation. In a more complex example, a viewer looking for a text of Othello will be able to choose between a version based on the Quarto, a version based on the Folio, or a version with readings from each, highlighted in different colours to distinguish them. In each case the "base" text will be the same, so that any changes will need to be made in only one place.
If there is a moral to this tale, I would argue that if politics is too important to be left to the politicians, site design on the Internet is too important to be left to the designers and the programmers. If it is expressed with sufficient force, a counter-ideology can be sustained in at least a corner of the communication space that is the Internet. This counter-ideology might establish as a primary objective the creation of sites that allow for maximum flexibility in viewing the materials; we should not be so puritanical as to reject the attraction of visual rhetoric, the pleasure of ornament or the power of multimedia to extend communication beyond the word, but, as in other civil forms of communication, the rhetoric should enrich the content, and should respect the viewer. [16] We should exercise what influence we have as users by being cranky and sending notes -- firm if politely unincendiary -- to web designers of sites that restrict our choices.
Those of us developing materials for web sites should insist on intelligent tagging: relative font/table/frame descriptions rather than fixed, and logical rather than descriptive HTML wherever applicable; we should generate as many literary texts as possible in some form of the more fully conceptual meta-language SGML (Standard Generalized Markup Language), while at the same time being aware of the fact that SGML as used in the Humanities has been developed to allow for the intelligent tagging of printed books and written manuscripts, in its own way perpetuating some structures that are anachronistic in the electronic text. [Hypernote 3]
Much of our energy in the Humanities will of course be directed towards representing the printed page on the electronic screen, as we provide machine readable editions of standard works. And as academics working within a tradition that has made a virtue of preserving the past, this is as it should be. But if we wish fully to humanize the new technology of the Internet, we must make every attempt to discover how to use the new medium in ways that exploit its dynamic text creatively, recreating and transforming the past rather than merely preserving it. We need to learn how to write and design for the electronic world, taking advantage of its capacity to be both linear and lateral, fixed and contingent; we need to ensure that we do not require our sites to wear the hand-me-down technology of print; we need to set an alternative example to those who seek to chain the pixel to paper.
[H-2] In HTML, as extended by Netscape and Internet Explorer, various attributes of a table cell or frame can either be expressed in relative or absolute terms, as can font sizes. The available attributes of a frame are particularly inviting for those who wish to impose a single way of viewing material: not only can frames be defined in size by pixel, but the <frame> tag includes the accurately named attribute <noresize>.
The increasing number of sites that offer materials in Adobe Acrobat format are a further indication of the desire to transfer the page rather than to design the screen. In addition, the much-vaunted introduction of style sheets to HTML is likely to offer the designer even more power over the appearance and font faces of electronic pages. Inherited from the meta-language SGML (Standard Generalized Markup Language), style sheets are one of the many features of that language that are designed to provide consistency across platforms, including the most restrictive platform of all, the printed page. The debate between rigid and flexible display continues in the development of the specifications for style sheets: one goal promoted is that they "[support] presentation hints, not commands (make no guarantees)". [17] However, the draft specifications include the capacity for defining "spatial" properties as well as "relative" properties in many elements, such as margins, indents, and font faces, thus once again providing programmers with the capacity to force control over the user. On SGML, see the next hypernote.
[H-3] The specific guidelines within SGML developed by the Text Encoding Initiative [18] have become a standard in the encoding of literary texts, but there have recently been some cogent criticisms of them, even in the area where they are most effective, the tagging of printed works. Ian Lancashire has written of the difficulties of adapting SGML for the process of tagging Renaissance printed texts accurately; Lancashire remarks that "SGML and TEI make anachronistic assumptions about text that fly in the face of the cumulative scholarship of the humanities". [19] Approaching SGML from the point of view of an encoder working on a modern printed work, Michael Neuman shows that the encoder continually encounters problems of interpretation as the codes become more detailed, yet it is precisely the detail of the tagging that makes the text useful for analysis. [20]
To these shortcomings of SGML/TEI I would add a further one that becomes pressing in the realm of texts created specifically in electronic format. Paradoxically, while the TEI Guidelines in general avoid tags that describe the physical nature of the book, they assume a general model of a conceptual book that is shaped by the traditional printed medium, divided into <front>, <body>, <back>, and various divisions and subdivisions. This sequential and hierarchical ordering of text is ultimately inadequate in the electronic medium, where sequence becomes multi-linear rather than linear, and where there are often multiple and overlapping hierarchies. Jay Bolter writes perceptively of the inherent clash between hierarchy and association: "A hierarchy is always an attempt to impose rigid order upon verbal ideas that are always prone to subvert that order" (Bolter 1991: 22).
[5] A number of educational sites make the background image for their page a simulation of a notebook; see for example the pages for the Abbotsford School District (B.C.) at <http://www.sd34.abbotsford.bc.ca/>, and York University at <http://www.yorku.ca/>.
[6] From <http://www.microsoft.com/truetype/web/designer/face3.htm> (23 May 1997).
[7] HUMANIST discussion group (7 February 1997).
[8] Analysis can range from the algorithms employed by search engines on the Internet to the complex processes of textual analysis under development in Humanities computing, exemplified by such analytical programs as TACT, developed by Ian Lancashire, in collaboration with John Bradley, Willard McCarty, Michael Stairs, and T.R. Wooldridge (Lancashire et al. 1996).
[9] On conceptual versus physical markup, see Coombs, Renear, & DeRose 1993; see also DeRose 1993.
[10] A particularly interesting example is the doublethink evident on the site for Intel that provides information to developers on the new technologies becoming available for networking. Intel prides itself on developing chips that are powerful enough to sustain operating systems capable of multitasking and "multithreading" -- but it designs its screen so that only one window will be visible. See <http://www.intel.com/iaweb/exptech/index.htm>.
[11] Located at <http://www.cocacola.com/> (9 September 1997). The graphic text reads "A Coke is a Coke no matter Where on the planet you drink it. but a Coke light can be a diet Coke. and a mello Yello can be a Lychee Mello. Fanta isa dozen different things -- Peach in Botswana, passion fruit (what else?) in France, and flower flavored in Japan (huh?) Other countries have their own flavors -- only Italy can pour a Beverly (and some travelers who've tried it are just fine with that)."
[12] The WebTV home page is at <http://www.webtv.net/> (22 May 1997).
[13] From the WebTV page. Clicking on the image of the remote control brings up a new window with a full size graphic, complete with Java applets that explain what each control will do.
[14] The US Globe site is at <http://www.shakespeare.uiuc.edu/>; (22 May 1997 [since revised]).
[15] The home page for Netscape has a JavaScript applet that is disarmingly called "sniffer".
[16] There are, of course, many others arguing for sensitive and flexible Web design. Kathy E. Gill is exemplary; in a posting on HUMANIST (8 February 1997) she wrote: "There are many resources 'out there' for critiquing web sites -- my philosophy of design and ease-of-use can be found at http://www.enetdigest.com/design/design.html -- I publish a weekly guide/critique to web sites. I have links to other resources as well -- more at http://www.dotparagon.com/design.html."
[17] See "Cascading Style Sheets: a draft specification" at <http://www.pku.edu.cn/on_line/w3html/Style/css/draft.html>.
[18] In print media see Goldfarb 1990; also Sperberg-McQueen & Burnard 1994. In electronic media see "TEI Guidelines for Electronic Text Encoding and Interchange (P3)" at <http://www.hti.umich.edu/docs/TEI/>; "The official TEI Home Page" at <http://www-tei.uic.edu/orgs/tei/>; "A Gentle Introduction to SGML" at <http://ota.ahds.ac.uk/teip3sg/>.
[19] Lancashire 1995. See also Ian Lancashire's RET Encoding Guidelines <http://library.utoronto.ca/www/utel/ret/guidelines0.html>; the home page for RET is found at <http://library.utoronto.ca/www/utel/ret/ret.html>.
[20] See Neuman 1995. For an introductory discussion of the advantages and difficulties of the related process of "lemmatizing" -- marking texts specifically for searching and developing concordances -- see Siemens 1996.