A Bookmaker's Workbench - CiteSeerX

5 downloads 244 Views 533KB Size Report
created from PDF or HTML files using a shell script or web service, it is not so easy for ... method, margin size, typeface, font size, colors, style of illustrations and ...
A Bookmaker’s Workbench Veronica Liesaputra and Ian H. Witten Department of Computer Science University of Waikato Hamilton, New Zealand {vl6, ihw}@cs.waikato.ac.nz designed to ensure an attractive product whose content can easily be assimilated and whose readers can easily find information and navigate around.

ABSTRACT

We have been developing electronic Realistic Books that combine the natural advantages of electronic documents— full-text search, hyperlinks, animation, multimedia—with those of conventional books—the ambient information provided by the physical object, analog page turning, random-access navigation, bookmarks, highlighting and annotation. Although simple Realistic Books can easily be created from PDF or HTML files using a shell script or web service, it is not so easy for book designers to take advantage of advanced features that are not normally represented in the input files. This paper describes the Bookmaker’s Workbench, an interactive system intended to help book designers produce Realistic Books. It incorporates many features, including a text mining option that automatically identifies significant key terms and marks them visually in the text, the ability to incorporate synonyms automatically into the full-text search capability, and include automatically generated back-of-the-book index. A user evaluation is reported that demonstrates the system’s usability and learnability.

We have been developing electronic Realistic Books that combine the natural advantages of electronic documents— searching, hyperlinks, animation, multimedia—with those of conventional books—the ambient information provided by the physical object, analog page turning, random-access navigation, bookmarks, highlighting, and annotation [6]. The features of conventional books have developed and improved over centuries; our aim is to replicate and enhance them. The Realistic Book system uses Adobe Flash as the presentation device. Studies of reader performance while searching and browsing show that participants perform tasks significantly quicker using Realistic Books than with printed books, HTML and PDF formats, without any loss in accuracy [5, 6]. Simple Realistic Books can easily be created from PDF or HMTL files using a shell script, or a web service that we provide. However, to take advantage of advanced features that are not normally present in such files, book producers must edit a configuration file manually or work directly in the Flash application—and both require specialist knowledge. To rectify this we have designed, implemented, and evaluated an interactive Bookmaker’s Workbench to facilitate the process of making Realistic Books.

Author Keywords

Electronic book, Book editor, Flash application. ACM Classification Keywords

H.3.7 [Information Storage and Retrieval]: Digital libraries, User issues; H.3.7. [Information Interfaces and Presentation]: User interfaces, Evaluation/methodology, Graphical user interfaces, Interaction style

The conventional publication process for printed books has three stages: editorial, design and production [9]. In the editorial stage, copy editors help finalize the book’s content. In the design stage, the structure and visual aspects of the book are determined. The book designer chooses details like page size and format, cover design and binding method, margin size, typeface, font size, colors, style of illustrations and tables—anything that affects how the page looks. In the production stage, the table of contents, subject index and bibliography are inserted and the manuscript is formatted according to the designer’s specification. Proofreaders read each page carefully and examine its layout in a cycle of checking and correction; then the proofs are handed off to the printer. Self-publishing authors often perform one or more of these tasks themselves.

General Terms

Design, Experimentation, Human Factors, Performance INTRODUCTION

Publishing a printed book is a major undertaking. Each of many components—size, format, binding, page layout, placement of chapters and sections, tables of contents and figures, subject index, and so on—must be carefully Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, or republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee. CHINZ 2011, July 4–5, 2011, Hamilton, New Zealand. Copyright 2011 ACM 978-1-4503-0676-8/11/07...$10.00.

The process of producing high-quality electronic books involves much the same steps. Furthermore, for online reading to be generally acceptable, electronic book

1

Figure 1. Bookmaker’s Workbench interface

applications must provide features that increase reader performance by transcending the affordances of paper books [1]. Of course, such applications will only find wide usage if it is easy to convert conventional computerreadable documents into the special form they require [2]. This is the role of the Bookmaker’s Workbench.

FEATURES OF REALISTIC BOOKS

The reader’s view of a Realistic Book shows a double-page spread flanked by a stack of page edges to the left and right that may include bookmark tabs [6]. Pages can be turned by ―grabbing‖ them with the mouse, or simply clicking them; clicking a bookmark goes directly to the target page. A toolbar appears beneath the double-page spread that gives the reader access to forward and back arrows, selection by page number, animation controls for pages with overlays, a magnification device, annotation tools, and a search box. The book designer can suppress individual tools or the entire toolbar; readers can suppress the latter. To the left of the double-page spread appears a preview area that gives a thumbnail view of the book; again, this can be suppressed by the reader.

Users interact with the Workbench to transform electronic documents into Realistic Books. Implemented within the Adobe Flash environment, it communicates with a PHP application server to save the books and any changes made to them. It addresses the design and production stages, but not the editorial processes. This paper uses the term book designer to denote the various players involved during the last two phases of the publishing operation, such as typesetters, indexers and proofreaders.

To visualize all this, see the Bookmaker’s Workbench in Figure 1, of which the lower 80% gives a preview of the book that is being constructed in a format that is identical to the reader’s view described above. Above this is the repository area, which holds the raw material from which the book will be created—all the video, audio, text and image files that the designer has uploaded. Beneath the repository area is the designer’s toolbar, from which the

A Realistic Book is constructed from one or more electronic files that the designer uploads. Its logical structure, layout, content, physical properties, and reader services are all defined in a template file. Sections 3 and 4 explain how a template file for the book is automatically translated into the book itself. Section 5 presents the evaluation study procedure and results.

2

Parameter

 running headers  automatically assigned section numbers and figure numbers  hyperlinked Table of Contents and Table of Figures  bookmarks for sections and figures  user-defined bookmarks  tools for annotation, including textual and freeform notes, and highlighting  an incremental full-text search feature that records the location of search hits by creating temporary bookmarks  content enrichment via automatically-generated hyperlinks  hyperlinked Subject Index  a synonym resource that is used during full-text search.

Description

Pages: PageWidth PageHeight PageType

width of a page height of a page type of papers used for the book’s pages PageColour base color of the book’s LeftMargin left margin of the page RightMargin right margin of the page TopMargin top margin of the page BottomMargin bottom margin of the page HeaderHeight font size of the running header PageNumHeight font size of the page number Covers: CoverType type of papers used for the book’s covers CoverColour base color of the book’s cover BookMargin distance of the page from the cover Page edges: xOffset horizontal offset between each paper yOffset vertical offset between each paper Bookmarks: BaseColour default color of a bookmark TabStyle styles of the bookmark SnapTo bookmark targets Page flipping: FrontPause pause time before turning the front cover BackPause pause time before turning the back cover TurnPause pause time before turning the book’s pages TurnSpeed speed of turning a page General: ZoomStyle magnification level of a book ShadowLevel darkness level of the shadow

In addition, there is a substantial amount of metadata that controls the appearance of the book, as shown in Table 1. Most of these features are associated with the book itself, but some are stored on a reader-by-reader basis (e.g., userdefined bookmarks and annotations). The last three features in the above list are particularly noteworthy. People often read documents extensively, first determining whether a book is worth reading and then focusing on portions of the text in order to locate the desired information [3]. Readers readily notice headings, illustrations, charts and tables, all of which stand out visually. They also scan the text for key words and phrases. To facilitate this, the Bookmaker’s Workbench is able to automatically identify significant key terms in the text and mark them visually. Wikipedia is used as a comprehensive knowledge base for this operation, and the visual mark takes the form of a hyperlink to the relevant Wikipedia article. Furthermore, when readers move the mouse over a term, a short description of it pops up. For example, Figure 2 shows part of a book page containing a news story about the day President Obama took his oath of office, which has been automatically augmented with links to relevant Wikipedia articles: Barack Obama, civil rights, Martin Luther King and African Americans, amongst others. In this book hyperlinks are underlined and colored blue. The mouse is hovering over the phrase 44th President of the United States of America, which brings up a popup showing an excerpt from the relevant Wikipedia article.

Table 1. Modifiable parameters for visualizing a book

book designer can customize visual aspects of the book such as its dimensions, fonts and page layout, and select suitable reading services that it will incorporate.

All this information is generated using the text mining technology described in [8]. This automatic identification of key terms also allows a hyperlinked Subject Index to be created automatically by collating them together with hyperlinked references to the appropriate page (and page number).

Realistic Books can have many features, and the designer must specify which ones are to be incorporated into the actual book that is being produced. Features include  cover image and color  page content: PDF files, HTML files, image files, video and sound files  different versions or ―overlays‖ of each page, selected under reader control (like animated PowerPoint)  pagination of textual matter  page numbering, including separately numbered front matter

Previous research has shown that one reason why people fail to locate the information they seek is that they describe it in terms that differ from the terminology used in the document [4]. One solution is to increase the number of ways in which each piece of information can be denoted. The Workbench generates a synonym table automatically as

3

Button name

Icon

upload add bookmark add section bookmark list section list hyperlink options document tools

Figure 2. A news story excerpt with links to Wikipedia articles

layout repaginate

a byproduct of the above-mentioned text mining operation. This table is consulted automatically during full-text search to increase the chance of getting a hit, even when the reader uses different terminology to that adopted in the book.

page break text format save

In Figure 2, President Obama, President Barack Hussein Obama and 44th President of the United States of America are all linked to the article Barack Obama. Thus these four terms are considered to be synonyms. Although the phrase Barack Obama does not appear in the text, it is inserted into the synonym table because Wikipedia uses it as the term for the topic (along with any synonyms, or ―redirects‖, that are defined in Wikipedia). When a reader enters any of these terms into the find box, all locations of the term and all of its synonyms are returned—creating semantic links that connect related passages that contain information expressed in different terms. The search algorithm operates incrementally, so that hits appear as soon as the first character is typed and are refined with each successive letter.

Table 2. Buttons in the designer toolbar

the book. When a PDF file is uploaded, each page is saved as a page image in Adobe Flash format (.swf). Similarly, audio or video files that are recognized by Adobe Flash are wrapped into a media player and stored as swf files. All these page images are displayed in the repository area shown at the top of Figure 1, with names such as page_28.swf and page_29.swf that are assigned by the system. The template file

As noted earlier, the complete structure of a book is specified in a template file. It is marked with a red star in the repository area. Any text document couched in a standard markup language such as XML or HTML (or even a flat ASCII file) can serve as the book’s template. Designers run a conversion script on HTML or ASCII files to convert them to XHTML format. If no template is specified, one is automatically generated. By default, it will create a hardcover book and use all the files in the repository area as the contents. This is a common way of generating an initial version of a book, which will then be customized.

CREATING REALISTIC BOOKS

Realistic Books are generated under the control of a template file written in XHTML. This file defines the sequence of pages that constitutes the book and the services to be provided to readers, along with any associated metadata such as page size, margins, title and where the main text starts (Table 1). It may also contain text, with HTML markup, that will go into the book. There are three phases: acquisition, automated document processing, and customization. The buttons in the designer toolbar are shown in Table 2. The book itself does not exist (and therefore does not appear on the display of Figure 1) until it has been generated during the Automated document processing operation described below. A common way of proceeding is to first make an initial version of the book and then customize it by adding new content to existing pages, and adding new pages.

If a book designer chooses to work with an existing book, all files specified in its template file are automatically uploaded to the repository area. Automated document processing

The system paginates any text in the template file, or in any other text documents that have been uploaded, using the algorithm described in Section 4. Link information is extracted and saved for later use.

Acquiring source material

Designers use the upload button (Table 2) to add to the repository area every file that contains source material for

4

Then the structure of the book is determined. Sections marked by
tags are assigned a number and level (e.g., 1., 1.2., 7.1.9); level-1 sections receive special treatment (they are called ―chapters‖ in the present article). Similarly, images are numbered according to their enclosing section identifier and order of appearance (e.g., Figure 1.1, Figure 4.1.3). A Table of Contents and List of Figures are generated from the titles of the different sections and figures in the book respectively. Bookmark tabs are added to the beginning of sections and figures. Running headers and page numbers are also created automatically.

during the pagination process. If necessary, book designers can make further modifications manually. First, a book’s template file is parsed and the text is formatted according to the style described in the template— for example, section titles are given an appropriate typeface and bookmark tabs are attached to the relevant pages. Link source and destination are stored in the hyperlinks table, both for hyperlinks that appeared in the original document and ones added by the text mining processes. Line and page breaks are created according to the guidelines explained below. At the end of the pagination stage, the Table of Contents, List of Figures, and back of the book Subject Index are automatically generated, each entry being expressed as an internal hyperlink, and inserted into the appropriate position in the document.

Predefined design styles are applied to particular elements of the book. For example, section titles are given a boldface style with a large font size that depends on the level of the section. Unwanted tools in the reader tools area are disabled. Each occurrence of every word in the document is processed and stored in a full-text index.

Margins frame the content of a page. The objective of margin design is to enhance the book’s utility and ensure that every pair of facing pages produces a pleasing aesthetic effect when the bound book lies open. Margins normally occupy up to 40% of the page, and the organization of the text is often adjusted to prevent unacceptable situations arising during pagination. The ContentWidth and ContentHeight of a page are equal to the size of the page minus the size of the page margin and the size of the page number and running header. The body text will never expand outside this region.

As mentioned in the previous section, text mining techniques are applied to determine keywords and key phrases in the running text. This results in the addition of Wikipedia hyperlinks and explanatory pop-ups within the text, an automatically-generated Subject Index, and a synonym resource for full-text search. Customization

Once the document has been processed as described above, a Realistic Book is generated. The appearance of each component in the book space, reader tools areas and preview area are updated accordingly.

A page break is introduced before the beginning of each chapter, and every chapter begins on a right-hand (oddnumbered) page. If the end of a chapter falls on an oddnumbered page, a blank page is inserted to ensure that the next chapter begins on an odd-numbered page.

Designers can further customize the resulting book. The designer toolbar contains a set of functions for modifying the book’s appearance and the reading services associated with it—for instance, redefining its logical structure, associating different styles with section headers, and changing the page margins. After these changes, the template file is updated accordingly and the document is reprocessed according to the new specification. The cycle is repeated until the designer is satisfied with the appearance of each component of the Realistic Book.

When a paragraph does not fit on to a page, orphan and widow rules are applied. Section titles are always located on the same page as the section’s first paragraph. A paragraph is moved to a new page when fewer than OrphanCount lines remain on the current one or when fewer than WidowCount lines would flow on to the next page. All pictures are automatically scaled so that their width and height do not exceed ContentWidth and ContentHeight respectively. A zoom facility allows readers to view images at their original size, to solve the problem of poor legibility at reduced sizes. If insufficient space remains on the current page for a figure, it is placed on the next one. Figure captions appear on the same page as the illustration.

PAGINATION

Pagination is a vital step in the conversion of electronic text formats such as HTML or plain text into a book, and is performed automatically by the Bookmarker’s Workbench. It is defined as the process of laying out parts of a document into pages [10]. Of course, text cannot be divided into pages willy-nilly: for satisfactory results there are many typographical constraints that must be respected [7].

In Adobe Flash, the same piece of text may be rendered on different occasions with slightly varying space and letter size (for unknown reasons). This means that the location of line breaks can change when the document is reloaded, altering the paragraph’s height and requiring the entire book to be repaginated to ensure that the text does not overflow its allocated space. This is highly undesirable, because the pagination process is expensive; also, it gives books a very unstable appearance. Consequently the Bookmaker’s

Finding an optimal pagination for a document that satisfies a formal specification of constraints for the criteria described in [7] is computationally expensive, although heuristic methods help greatly. However, implementing pagination algorithms is not the focus of this research, so the Bookmaker’s Workbench applies just a few basic rules

5

Workbench inserts explicit line breaks into the text to guarantee the same paragraph size whenever the document is reloaded.

editing environment. We assessed the usability of each method based on learnability, efficiency, effectiveness and user experience.

Extra space within words can degrade the quality of text presentation and distract the reader’s attention. In the Workbench, words are not hyphenated across line or page breaks. Successive words in a paragraph are inserted into a line until the text width exceeds the content’s width. When a word does not fit, the line is broken and that word begins the next line. Because an explicit line break is inserted after each line, the text is unjustified. This has the side effect of avoiding rivers of space, because the space between words cannot exceed the space between lines.

Setting an interactive GUI against an HTML editor is something of a straw-man comparison, although the former might be frustratingly slow because it frequently repaginates in order to show a faithful preview of the book. Also, the learnability of each system is only informally assessed, because the experimenter gave a verbal explanation and demonstration of the two systems. Consequently we are more interested in qualitative comparisons and user comments on the methods than in quantitative results.

During the line breaking process, full-text indexing is performed. All words in the book—and all their synonyms—are indexed and stored in the index table. Index terms are case-folded; however, stemming is not performed. We have found that the ―redirects‖ that have been recorded manually in Wikipedia provide a far more reliable basis for full-text searching than any automatic stemming algorithm, and as noted in Section 2 these are incorporated into the synonym table and used during full-text search.

Participants

We recruited 16 high school and university students aged 15–40 from a variety of disciplines to create two books using each method. Participants had used computers extensively for several years and had, as readers, encountered Realistic Books at most twice before. None had previously used the Workbench, but all knew how to create a web-page using Dreamweaver. Procedure

During pagination, any hyperlinks that occur in the manuscript are interpreted and recorded. Designers can create a link to a particular page in the book or to an external source. Figures, chapters, paragraphs, and sections can all be referred to by unique link identifiers rather than by page number, to cater for the dynamically paginated nature of Realistic Books.

Participants came individually to the study and were asked to complete a profiling questionnaire that recorded age range, gender, language proficiency, and their experience with computers. Each participant was first trained to use both methods to create a book. After gaining confidence, they performed two tasks, which they undertook in the same order. Each task asked participants to fill in missing information in a 50-page book—e.g., insert a bookmark to the page that mentions a specified term, add an image that best describes this text—and to modify its appearance—e.g., change the book’s size to make it more readable. This involved a total of 15 subtasks.

Whenever the book is repaginated, the page numbers of link destinations may change but the link identifiers remain the same. During the pagination process, every link identifier and its associated page number are stored in a table. Whenever a reader clicks an internal link, the hyperlink table is consulted to determine the appropriate page number and the book is opened at that page. Whenever the book is repaginated, the table is updated. Designers do not need to manually update link destinations unless they refer explicitly to page numbers.

Participants used the Bookmaker’s Workbench for one task and the HTML editor for the other, but in different orders Task 1 using the Workbench and Task 2 with the text editor, and vice versa. The 16 possible orderings were allocated evenly between participants, so that eight created each book using the same tool.

EVALUATION

The final outcome of the book creation process is a file called book.htm that contains all the relevant metadata and defines the book’s content in terms of the file names of each page image. An alternative, more primitive, way of creating books is to use an HTML text editor to construct this file directly, and shell scripts exist to convert other documents such as PDF files to page images for the book.

The functions participants used during each task and the time they took to learn each tool and to finish each task were recorded. Participants were encouraged to ―think out loud‖ and make comments. Having finished both tasks participants completed a questionnaire to capture their opinions. They had to choose which method was useful for them, easier to use, more pleasant and engaging to use, and finally which they preferred overall. They were asked to explain their choices and state which features they liked and disliked, as well as suggest improvements.

In order to gain some experience with how users perceive the Bookmaker’s Workbench a small evaluation study was performed. The study compared user experiences with the Bookmaker’s Workbench with the more primitive method. Adobe Dreamweaver was used as the HTML editor. It supports HTML tag auto-completion and provides a visual

6

opening tags in the correct placed in the document. In fact, every time users made any kind of change to the book template file, however small, they would straight away try to visualize the resulting Realistic Books to check that the book template file was still well-formed.

Results

We evaluated performance based on the outcome and process measures defined above. Learnability

Quantitatively, the learnability of a system was measured by the time it took subjects to build a book, and the number of questions asked during the learning process. It is assumed that a system is easier to learn and understand if subjects spend less time and ask fewer questions during the training process. Subjects were also asked at the end of the study whether they found the system easy to learn.

Subjects did not want to run the conversion script to automatically ensure that the template file is valid, because the script took quite some time to finish. They only used it when they had made extensive changes. The Workbench gives an instant preview of the book every time designers make any changes, and ensures that the template file is well-formed. Participants only needed to concentrate on completing the task. Some spent longer creating a book using the Workbench than the HTML editor only because they performed extra measures that were not required to complete any sub-tasks to heighten the aesthetics of the finished book with a variety of information (i.e. graphics, audio and video).

The time spent creating a book using the HTML editor ranged from 18 to 45 minutes, with an average of 37 minutes. The average number of questions asked was 15. With the Workbench, the time taken ranged from 10 to 40 minutes, with an average of 28 minutes. On average, 9 questions were asked during the training process. A t-test analysis between the two systems on both criteria—showed that the difference between the systems was statistically significant at the five percent level.

Effectiveness

With the Workbench, participants could clearly see the correct section level and position in the page. However, a few forgot to press the add section button in the designer tools area when they were adding a new section. They typed the title straight into the page’s main text area, so the section was not added to the automatically generated table of contents and the page was not bookmarked as a section.

Regardless of their computer experience, participants found the Workbench to be natural and needed less explanation about its usage than the HTML editor. They were able to quickly discover and learn the tools required to complete a task by themselves without consulting the user manual or the researcher. When asked, all participants commented that the icons chosen to represent the tools were clear and they could guess the function of each button in the toolbar. With the HTML editor, participants relied heavily on the user manual to know what tags were appropriate for completing a task.

Because the section number was not visible when the book template was viewed with the HTML editor, when participants were asked to insert a new section at the specified location, most either specified an incorrect section level or inserted the section tag at the wrong place.

Efficiency

All participants made fewer errors when creating books with the Workbench than with the HTML editor. The average number of sub-tasks that were completed successfully was 67% using the HTML editor and 87% using the Workbench. A t-test analysis between the two systems showed that the difference was significant at the five percent level.

After just one training session, lasting less than 50 minutes on each system, most subjects were able to create a book with minimal requirements (i.e., front and back covers, table of contents and content comprising both text and graphics) independently, without any help from the researcher, in about 20 minutes less than during training. This could be interpreted as a positive indication that subjects easily understood how to use the program.

User experience

Having completed all tasks participants were asked which tools were most useful, easy to use, pleasant and engaging, and preferred overall.

The average times spent per sub-task ranged from 10 seconds to 9 minutes with the HTML editor, and from 5 seconds to 5 minutes with the Workbench. There was little variation in the time between participants. A t-test analysis between the two systems showed that the difference was significant at the five percent level.

Most subjects (88%) commented that while the HTML editor suffices for short documents without any sections or animation, they preferred the Workbench for long and more complex books. With the Workbench, they did not need to know the tags required to specify and control the appearance and the contents of the book. They could rapidly start doing their work instead of spending time ensuring that they had entered the right tags or that the template file was a well-formed XHTML file.

Most participants completed all sub-tasks faster using the Workbench than with the HTML editor. It was observed that all participants spent most of their time in the HTML editor making sure that the book template file was a valid XHTML file. For example, to change the format of a specified piece of text in the book, participants needed to ensure that they had inserted the appropriate closing and

7

With the Workbench, they felt that they would still know how to use it after a time. They do not have to re-learn it all over again. Because it provides a preview of the final product, they felt more confident in exploring each tool that was available. They could see the results of the changes they made, instantly. With the HTML editor, participants felt that they always needed to consult the user manual to create a book.

convenience. This paper has explained the two stages involved in this transformation: automated document processing and customization. First, text is paginated and can be augmented with automatically-generated hyperlinks to an external knowledge source (Wikipedia). Then, the designer can add various features to the document, which may involve repagination, and provides readers with suitable reading tools to perform their tasks.

12% of participants said that they did not have any preference about which program to use. They felt that once familiar with the tags, they would not need to rely on the book preview to make any changes to the appearance and contents of the book, nor the list of reader tools available to the book readers.

A user study has shown that people can produce books more readily with the Workbench than with a plain HTML text editor. ACKNOWLEDGMENTS

We acknowledge the entire New Zealand Digital Library Project team for their unstinting work in providing an environment that makes this kind of research meaningful— and fun. This work is funded in part by Google.

Most subjects criticized the size of the Workbench’s book space area: they wanted more space for the book. They suggested that there should be a button that allows users to minimize or hide the repository area and the design toolbar area. They also wanted a more explicit way to add additional synonym terms to the synonym table.

REFERENCES

Participants said that when learning how to use the tool, automatic repagination after each change was useful because they could see the effect straightaway. However, for long documents, repagination takes some time to finish. Participants suggested that there should be an option to turn off automatic repagination and let the designer decide when repagination should be performed. For example, when trying different page and cover sizes of the book, designers did not want the text to be repaginated. They only wanted to repaginate once they were satisfied with the chosen page and cover size. Users also suggested that there should be a list of possible front covers that designers could choose for their books, as well as color panel that allows them to select a color instead of entering a value. CONCLUSION

The physical design of printed books provides many cues that help readers to find their way around the text. Conversely, electronic documents have their own advantages of accessibility, searchability, hyperlinks, multimedia material, and the possibility of automatically enriching the content through text mining. Realistic Books have been developed to combine the advantages of the two. However, constructing a book is a major undertaking. The Bookmaker’s Workbench allows designers to transform electronic documents into Realistic Books, perhaps bringing together different sources of material and including various accessibility features for the reader’s

1.

Dillon, A., Richardson, J., and McKnight, C. (1989) ―Human factors of journal usage and design of electronic texts.‖ Interacting with Computers, 1(2), 183–189.

2.

Egan, D., Remde, J., Gomez, L., Landauer, T., Eberhardt, J., and Lochbaum, C. (1989) ―Formative design-evaluation of superbook.‖ ACM Trans. Information Systems, 7(1), 30–57.

3.

Fischer, S. R. (2004) A history of reading. London: Reaktion Books.

4.

Furnas, G. (1985) ―Experience with an adaptive indexing scheme.‖ Proc. of CHI’85 Human Factors in Computer Systems, 131–135.

5.

Liesaputra, V. and Witten, I.H. (2008) ―Seeking information in realistic books: A user study.‖ Proc. Joint Conf. of Digital Libraries, pp. 29-38.

6.

Liesaputra, V., Witten, I.H. and Bainbridge, D. (2009) ―Creating and reading realistic books.‖ Computer, 42(2), 46–55.

7.

Martin, D. (1989) An outline of book design. London: Chapman and Hall.

8.

Milne, D., and Witten, I. (2008) ―Learning to link with Wikipedia.‖ Proc. ACM Conf. on Information and Knowledge Management, pp. 509–518.

9.

Mogel, L. (1996). Making it in book publishing. New York: Arco.

10. Rubinstein, R. (1988) Digital typography: an introduction to type and composition for computer system design. Reading: Addison-Wesley.

8