From Book to Screen: A Window on Renaissance Electronic Texts

Introduction: Text as Graphic

1. A good place to start is the obvious: a computer screen is nothing like a book. The book is a wonderfully rich source of stimulus; as we read we are guided by visual signals provided by the type and perhaps by images, and as we progress we receive the regular reward of turning pages, pages which have a satisfyingly three-dimensional existence, and which even smell good. The computer provides a two-dimensional screen that displays text and graphics in a very similar manner, but there is far less sensual feedback for the reader. Much of the energy that goes into the development and design of electronic texts is spent on recreating the kinds of expectations and rewards that readers of books are used to; the danger is that the new medium will become limited by its predecessor, and that the screen will simply be seen as a poor imitation of the original book.

2. Eventually there will no doubt be as many electronic editions of Renaissance texts as there now are on paper. They will serve different audiences, from the introductory to the scholarly; they will vary from largely unadorned, unformatted ASCII texts designed for statistical analysis to highly sophisticated interconnections of text, graphics, sound and video clips; and their means of delivery will vary from the Internet to CD ROM and possibly other methods as yet unthought of. Discussions of the future of electronic texts tend, naturally enough, to focus on the most spectacular ways in which the meaning of "text" can be extended to include sound and video components. In this paper, however, I want to begin at the beginning: to consider the way familiar words are actually displayed on the screen instead of a page. Taking the example of an ideal design for an electronic edition of Shakespeare, I want to look at the way the basic texts of the plays themselves, seen as pixels on the screen rather than words on the page, have to be treated as graphic images and require surprisingly complex decisions to be made about their appearance.

3. In a well-designed printed book the text is effectively transparent: readers scarcely notice the layout, the fonts used, the line breaks, or the page breaks. Though house styles differ in detail, the overall appearance of a page in any modern edition will follow well-established conventions. In a student edition of the collected Works (Riverside, Signet, Bevington, and so forth) the text will be in double columns with line numbers on the right, explanatory notes keyed to the lines at the bottom. Elsewhere there will be summary comments on the text with a list of major variants. It is an interesting comment on the effectiveness of early book design that apart from the addition of footnotes and improvements in typeface quality made possible by advances in technology the modern complete editions look surprisingly similar in layout to the First Folio (1623).

4. Single play editions differ in that the page is more generous in offering white space, and in more scholarly editions (Arden, New Cambridge) there is typically a fairly detailed collation, in very small print, between the text and the explanatory notes. The widest variation seems to be in speech headings (full names or abbreviated, small caps or italic, and so on).

5. On a computer screen all these design elements can be reproduced without change, effectively re-creating pages of a book on the screen. To do so, however, is to ignore the real differences between the media, since all these elements become potentially dynamic on the screen. Fonts and font sizes can be changed at will, and "page" becomes an inaccurate metaphor for the window in which the text appears; unlike a page the window can be resized. In fact, the designer of an electronic text faces a major problem in the degree to which the traditional interface of the book with its pages still dominates the way we expect the new medium to behave and how we expect to interact with it.

6. Shakespeare editions are especially interesting examples of material illustrating the rival claims of generic, conceptually tagged text designed to be interpreted by separately developed browsers, and graphic, object-oriented programming. The current evolution of HTML (Hypertext Markup Language) is an illustration: "good" HTML guides[1] stress the desirability of "conceptual" tags inherited from SGML (Standard Generalised Markup Language), and encourage the use of tags such as "<cite> </cite>" or "<em> </em>" which describe the content (citation) or the importance (emphasised) of the text rather than those like "<i> </i>" which describe the graphic appearance of the text (italic). Yet in recent months, especially with the appearance of the enhanced capabilities of a browser like Netscape,[2] sites on the Internet are increasingly exploiting graphic elements as a way of imposing a specific physical design rather than a generic, logical one.

7. A book after all is both an object and a container of concepts, and editors of Shakespeare are necessarily more aware of the physical object than many working with more recent or more stable texts: so much textual reconstruction has depended on the unravelling of the processes of making the original editions, from the graphic structures of Shakespeare's secretary's handwriting to the eccentricities of individual compositors, to the physical demands of creating a large book from movable type.[3] A valiant attempt to provide both the advantages of tagged text and adequate information about the physical book is the series under the general editorship of Ian Lancashire, Renaissance Electronic Texts, which attempts to "retain the original spelling, including sometimes many archaic letters, contracted or curtailed forms, marks of abbreviation, and ligatured typeface."[4]

8. It seems also that Shakespeare scholars, and even general readers, like to be reminded of the physical artifact of the original book. Thus for example the Voyager electronic edition of Macbeth[5] creates an attractive "parchment" look with its grey background framed in a kind of Celtic binding. In the discussion that follows I am assuming that the medium for the electronic edition is a CD ROM or equivalent rather than the Internet because the CD ROM offers the designer full opportunity to control the graphic appearance of the screen; the software that displays the images on a CD ROM is--or can be--included on the disk, and therefore is not subject to the same variability as a designer faces when dealing with multiple browsers.[6]

9. The theoretical implications of electronic texts and hypertexts have been extensively discussed in recent years,[7] but the actual number of applications using them is still limited, and the process of developing intuitive and reasonably universal conventions for hypertext and multimedia texts is still far less stable than equivalent conventions in the printed medium. There are many practical problems to be solved. The discussion that follows deliberately avoids reference to specific software packages or platforms, since it is addressing theoretical questions of design.

   Line Breaks in Prose

10. Theoretical discussions of hypertext have tended to stress its capacity for providing the user with non-linear, dynamic links between different blocks of text. The editor of a hypertext edition of Shakespeare has to deal with the problem that the basic text is unashamedly linear (we are told on good authority that the play will have a beginning, a middle, and an end). So linear is the text we have become accustomed to that we number its lines--and here we encounter a good example of the way the printed medium constrains the electronic. From the first quartos and the First Folio to the modern edition of choice, line breaks, especially in prose passages, are decided by the compositor, and all line references in the accompanying documentation are subsequently adjusted to those decisions. We are all familiar with the consequence: different editions have different line references, and the variation can be considerable by the end of a long prose scene.

11. The most familiar computer display of text for most of us will be the word processor, but it is important to realize that the screen we work with as we prepare our conference papers is to a perhaps surprising extent inherited from the printed page. When software programs enthusiastically trumpet that their design is WYSIWYG (What You See Is What You Get), what they really mean is What You See Is What You Want To Get On A Printout (WYSIWYWTGOAP). The ingenuity of the computer is employed in an attempt to make the screen as much like a book as possible: margins are fixed (unless we decide to print to a different page size and adjust them accordingly), and both page size and lineation remain constant if we resize the display window. We can, however, adjust the font size or the gap between lines in order to present the text more attractively. My students have discovered that they can make a short paper seem longer, or an overlong paper look shorter, by adjusting margins, fonts, and line spacing.

12. Text on the computer screen can, however, be more dynamic than this, and is often so in applications that do not expect the user to generate printed output. As the window is resized the text flows and rewraps to accommodate the new size, an arrangement that is arguably more native to the computer than the fixed recreation of a page. Chief among the advantages is the way that the dynamic resizing can make allowance for the increasing number of large monitors that are appearing on the market and on scholars' desktops; users should be able to choose a larger, more readable font size and adjust the window accordingly, or alternatively shrink the window so that other windows (containing indexes or graphics perhaps) could be made more visible. HTML allows some kinds of text to re-lineate and re-wrap as the window is resized, and browsers designed for SGML will do the same, depending on the kind of document that is being displayed.

13. It is in these electronic documents, arguably more fully adapted to the electronic medium than word processors, that the problem of cross-referencing arises. Hypertext linking within the document itself can be handled in such a way that the "lines" on the screen are irrelevant, but for the Shakespeare scholar there is a long tradition of line numbering in both verse and prose, so that by a process of accretion there are references in printed media to act, scene, line, that need to be referenced in some way. If lines of prose are to be dynamic and of variable length, the lines of many scenes in the plays will change according to the whim of the viewer.

14. There are several possible solutions.
    • 1. The editor/designer can reproduce a fixed page on the screen so that the numbering is no longer dynamic. Font, size and margins would be fixed. This approach denies the reader the full advantage of the electronic medium, especially if he or she has invested in one of the increasing variety of monitors larger than the standard 640 by 480 pixels. This is the approach taken by the Voyager Macbeth.

    • 2. Line breaks in prose could be fixed by making each of them a separate line defined by a "hard" carriage return. In this case, however, lines will wrap inelegantly in a small window, and will fail to fill out to a larger window. Prose in this form cannot be justified, so it will look more like verse than the equivalent passage in the printed medium.

    • 3. Line numbering can be anchored arbitrarily to a known printed text, with the numbers moving as the window is resized, even if the difference between line 100 and 110 is actually six or thirteen visual lines on the screen. This solution is not very different from the current convention of indicating at the start of a paper which printed edition line references will be taken from.

15. Whichever method is adopted, an electronic Shakespeare text will be to a degree constrained by the older print medium. There is in effect both input from and output to print: line references from the vast knowledge base of criticism work as a kind of input, and the need for a scholar to produce printable articles--or even electronic articles not dynamically linked to the text she or he is using--means that there is a potentially printed output that refers to specific, non-dynamic lines. The choice becomes one of accepting the medium and creating a new convention of cross-referencing passages, or imposing artificial line references of some kind.

    The Screen: Simple or Complex?

16. The physical division of the text on the screen into separate lines is a relatively simple illustration of the kinds of decisions that will have to be made about the electronic text as a graphic entity. More challenging will be the decisions to be made about the extent to which the visible presence of the editor should be perceived by the reader of an electronic edition.

17. Most modern editions indicate major editorial intervention, usually marking such passages with square brackets. Inserted stage directions or missing words inserted by the editor are examples. A slightly different use of the convention is followed in the way many editions treat the well-known instances where Shakespeare himself seems to have revised his text in such plays as Love's Labours Lost (5.2.817-22) by putting what is probably the first version in square brackets (angle brackets in the Riverside). The Bevington edition also uses square brackets to differentiate passages in King Lear that are limited to the Quarto.

18. The New Folger editions edited by Barbara Mowat and Paul Werstine provide a more elaborate example of a typeface that attempts to communicate the complexities and uncertainties that lie beneath the modern text. In the plays that have complex textual origins, the editors use not only square brackets but angle brackets < > and "broken" brackets in the modern text to indicate the provenance of various readings. A reader who has taken the trouble to read the textual introduction--and who remembers what the unfamiliar graphic elements signal--will immediately be aware which passages are unstable, and a naive reader will at least guess that something is going on under the surface. Even without editorial intervention of this kind, a modern edition, especially one of the complexity of the Arden or New Cambridge series, gives graphic signals that warn the reader of passages that are particularly problematic or difficult: as we turn the page we may see just two or three lines of text and a large slab of commentary, collation, or both.

19. The electronic edition, however, has the power both to present fuller documentation and to make the initial screen simpler. The ideal edition will include electronic and graphic versions of all significant original texts--quartos (if any) and the Folio--available at the click of the mouse for consultation and comparison; they will also provide a fuller and less runic collation, and full hypertextual commentary.

20. The particular case of Shakespeare texts provides a challenge to create an interface that does more than offer alternative sequential hypertext choices: to go beyond what Jim Rosenberg[8] calls "the confrontation with or" rather than with "and/and/and." The reader should be able on the one screen simultaneously to consult the modern text and the Folio and the Quarto and a table of contents and a navigational palette. In this situation the problems of intuitive management will be less those of navigation than of focus on a screen which may become cluttered with multiple windows: we might learn from high-end graphics and page-layout programs and use multiple floating windoids, like tool palettes, collapsible to their title-bars, or create new ways of stacking the screen with transparent or semi-transparent layered windows.[9]

21. But despite the wealth of annotation, collation, and commentary that an electronic edition can make available to the reader, the modernized text can be made to appear uncluttered, with the only indication of deeper layers of information and textual apparatus the menus and appropriate icons for accessing annotation. Alternatively the modern text could signal complexity by the same or similar means as the Folger editions. The decision that has to be taken on this issue is both a matter of graphic design and a choice that is in effect ideological: should the edition proclaim the complexity of selection and instability that lies behind any modern edition, or should it offer an apparently transparent top layer, assuming that the reader / user will be curious enough to explore the deeper layers and learn independently of the wealth of variance that is available? The designer's dilemma is in choosing between the typographical simplicity--even elegance--of an uncluttered screen which may make the text look deceptively authoritative and final, or a screen that implies directly through typographical means the challenge-ability of the surface text. The problem in the second case is the familiar one of signals crowding for attention with the effect of too many people speaking at once (as in the all too frequent overuse of STYLES and fonts in desktop publishing--though this attempt to reproduce the inelegance in a plain ASCII format is a rather feeble approximation of the real thing).[10]

22. Consider the extremely challenging King Lear, with its widely variant Quarto and Folio texts, and the continuing debate on the relationship between them.[11] In the case of a conflated text of King Lear it would be possible to mark, say, four kinds of passages: those in which Quarto and Folio are in essential agreement, those derived solely from the Quarto, those derived solely from the Folio, and those which involve editorial emendation of some kind. Rather than different kinds of brackets, colour could mark the different passages (it might be entertaining to consider the symbolic importance of the designer's choice of colours: black for uncontroversial text, green for the preferred text, red for the dangerous one, purple for the splendour of the editor's emendations). But different brackets and different colours have the same basic problem--they are not simple, and the graphic interface will seem cluttered or noisy. And busy though this text would be, there would still be a further need to indicate which passages are linked to commentary of various kinds, probably through icons in the margin rather than yet another signal generated from the typeface.

23. If we listen to the graphic designers and those who devise the interfaces for the software we are becoming increasingly attached to, the message is clear: keep it simple; keep it legible.[12]

24. In any case it is likely that the ideal design will include a user preference for showing the modern text with or without visual signals that communicate the textual choices made at any point. The scholar needing to know which sections came from which source could choose the less attractive but more informative screen; other readers could work with a cleaner text, the only indications of textual or subtextual complexity being the varying density of icons in the margin. An attractive solution might be to develop software which allowed the user to transform the text from simple to complex by holding down a modifier key in much the way that holding down the command and option keys on a Macintosh in HyperCard reveals the outlines of buttons; add the shift key and you see the fields as well.

    Conclusion

25. The editor of an electronic version of Shakespeare's plays has the burden of centuries of traditional printed text and commentary to carry into the new medium. Even in such basic design elements as the kind of windows and screens that will be used, some basic decisions have to be made in creating the conventions that will be readily understood by the modern user of the electronic text, and at the same time present that text in as accessible a fashion as possible. One way of rendering the new medium more intuitive to users bred on the book is to translate the electronic text into metaphoric fixed lines and pages, but to do so is to limit the power and flexibility of the computer screen. In the short term it may be most effective to develop methods of display that allow the user to express preferences, choosing to make the screen more familiar and book-like or more fully adapted to the electronic medium.

26. Because the editor of the electronic text is able to provide far more information beyond the text than the printed page, the resulting edition has the capacity to be something closer to a challenging "infinitive" rather than narrowing "definitive" edition.[13] The choices to be made in designing the windows for the modern text involve decisions not only about its purpose, but about the way its audience is to be treated: is the user to be visually reminded at all times of the complexity of the editing process and the instability of the modern text, or to be presented with a text that is (according to one's point of view) either elegantly uncluttered or deceptively transparent?

27. Even if we cannot yet give an electronic edition a satisfying tactile presence or a good smell, it can offer many things a printed edition cannot: a far greater amount of research material, and samples of graphic, sound, and video materials. It can also offer the user a power of choice over the appearance of the text. As editors and designers we must ultimately trust the reader who is looking at a simple surface text to be curious, to click and learn, to become an interactive user rather than a passive reader, and thus to realize that nothing is really simple. A book can have only one design, but as a graphic image the electronic text is capable of being transformed into many different designs, some pre-set, others customized. The reader must be given choice; we may have strong and well-documented opinions about the ideal combination of font and white space to promote readability, and we may want the surface text to be highly instructive, or sparse and uncluttered, but we are doing the new medium an injustice if we impose a single graphic, conceptual, or ideological structure upon it.

Notes

1. For an example of the discussion of what constitutes "good" HTML see James Tilton, "Composing Good HTML," 1995. URL: <http://www.willamette.edu/html-composition/strict-html.html>. "Tagged" text languages like SGML and HTML originated (and to a large extent remain) in the primitive world of ASCII text and line-oriented text editing and programming, but they include coded information that makes it possible for software to display the material in more sophisticated ways on a graphic user interface. In deliberately avoiding the physical representation of language, SGML and HTML are prime examples of WYSIANLWYG texts (What You See Is Absolutely Nothing Like What You Get).
2. Netscape Navigator. Copyright 1994-1995. Netscape Communications Corporation.
3. Randall McLeod neatly expresses the dialectic between object-driven and concept-driven electronic texts: in making a case for the importance of old typesetting rather than old spelling in reading the original texts of the Renaissance, he remarks, "the point . . . is simply whether material causality can dominate formal causality in this case." See "Spellbound" in Play-Texts in Old Spelling: Papers from the Glandon Conference, edited by G. B. Shand and Raymond C. Shady (New York: AMS Press, 1984, 89).
4. Ian Lancashire. RET Encoding Guidelines. December 1994. URL: <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>.
5. William Shakespeare, Macbeth, ed. A. R. Braunmuller (New York: Voyager, 1994). Introduction and commentary by David S. Rhodes; produced by Michael E. Cohen. System requirements: Macintosh with 25-MHz 68030 or better; System 7 or better; 3,500K of available RAM; 13" colour monitor; CD ROM drive.
6. Further advantages of a CD, at least for the present, are its capacity for reasonably rapid access to the large files needed for sound and video, and the opportunity that it offers for gaining copyright permission to reproduce materials that would not be permitted freely on the Internet.
7. Convenient links to works in this rapidly changing field are Jay Bolter's Writing Space: the Computer, Hypertext, and the History of Writing (Hillsdale, NJ: Lawrence Erlbaum Associates, 1991), George Landow, Hypertext, Hypermedia and Literary Studies (Cambridge, MA: MIT, 1991; edited with Paul Delaney; see particularly Larry Friedlander, "The Shakespeare Project: Experiments in Multimedia"), Hypertext: The Convergence of Contemporary Critical Theory and Technology (Baltimore, MD: Johns Hopkins UP, 1992), and Hyper / Text / Theory (Baltimore, MD: Johns Hopkins UP, 1994). A new edition of Jakob Nielsen's Hypertext and Hypermedia (Boston: Academic Press, 1990) is to be published this year.
8. "Navigating Nowhere / Hypertext Infrawhere." 1995. URL: <http://www.well.com/user/jer/NNHI.html>.
9. See for example the experimental approach taken by Beverly L. Harrison, Hiroshi Ishii, Kim J. Vicente, and William Buxton, "Transparent Layered User Interfaces: An Evaluation of a Display Design to Enhance Focused and Divided Attention." CHI '95 Electronic Proceedings. 1995. URL: <http://info.sigchi.acm.org/sigchi/chi95/Electronic/documnts/papers/blh_bdy.htm>.
10. A particlarly dreadful example of the interface shouting at the reader is the function "BLINK" in HTML.
11. A recent entry in an occasionally colourful debate is René Weis' King Lear: A Parallel Text Edition (London and New York: Longman, 1993), where "disintegrationists" appear and the arguments of "schools" of revisionists and conflationists are marshalled (1, 2).
12. In the field of book design, John Ryder offers cogent reasons for making legibility the most compelling principle of book design, and discusses the interconnection of font, line spacing, and white space--arguments that apply even more to the screen (The Case for Legibility [London: Bodley Head, 1979]). Some links to the subject of interface design are found in a quirky introduction to the subject, Bruce Tognazzini, TOG on Interface (Reading, MA: Addison-Wesley, 1992); more formal discussion is found in Harold Thimbleby, User Interface Design (New York: ACM Press, 1990). Both reinforce the need for the interface to present the user at least initially with a clear and limited range of choice, and both offer useful bibliographies. See also the regularly updated Human Interface Guidelines by Apple Computer (Addison-Wesley).
13. The term is taken from the stimulating article by "Random Cloud" [Randall McLeod], "The Marriage of Good and Bad Quartos," Shakespeare Quarterly 33:4 (1982), 422.

• Apple Computer, Inc. Macintosh Human Interface Guidelines Reading, MA: Addison-Wesley Developers P, 1995.


• Bolter, Jay. Writing Space: the Computer, Hypertext, and the History of Writing. Hillsdale, NJ: Lawrence Erlbaum Associates, 1991.


• Friedlander, Larry. "The Shakespeare Project: Experiments in Multimedia." Hypertext: The Convergence of Contemporary Critical Theory and Technology. Baltimore, MD: Johns Hopkins UP, 1992. 257-71.


• Friedlander, Larry. Hyper / Text / Theory. Baltimore, MD: Johns Hopkins UP, 1994.


• Harrison, Beverly L., Hiroshi Ishii, Kim J. Vicente, and William Buxton. "Transparent Layered User Interfaces: An Evaluation of a Display Design to Enhance Focused and Divided Attention." CHI '95 Electronic Proceedings. 1995


• Lancashire, Ian. Renaissance Electronic Texts Encoding Guidelines. December 1994. URL: <http://library.utoronto.ca/www/utel/ret/guidelines0.html>.


• Landow, George. Hypertext, Hypermedia and Literary Studies. ed. Paul Delaney. Cambridge, MA: MIT P, 1991.


• McLeod, Randall. "The Marriage of Good and Bad Quartos." Shakespeare Quarterly. 33:4 (1982): 422.


• McLeod, Randall. "Spellbound" Play-Texts in Old Spelling: Papers from the Glandon Conference. ed. G. B. Shand and Raymond C. Shady. New York: AMS P, 1984. 81-96.


• Nielsen, Jakob. Hypertext and Hypermedia. Boston: Academic P, 1990.


• Weis, René. King Lear: A Parallel Text Edition. London and New York: Longman, 1993.


• Ryder, John. The Case for Legibility. London: Bodley Head, 1979.


• Shakespeare, William. Macbeth. Ed. A. R. Braunmuller. New York: Voyager, 1994.


• Tilton, James, Carl Steadman and Tyler Jones. Web Weaving Reading, MA: Addison Wesley Developers P, 1995.


• Thimbleby, Harold. User Interface Design. New York: ACM P, 1990.


• Tognazzini, Bruce. TOG on Interface. Reading, MA: Addison-Wesley Developers P, 1992.

Return to EMLS 1.2 Table of Contents.