Multimedia Cd-Rom as a Medium for Manuscript Preservation and Dissemination: The Design and Development of "Treasures of Islam"

The paper discusses the use of electronic publishing for the preservation and dissemination of rare manuscript material. It is based both upon the authors’ earlier work on multimedia integration and upon their experience of designing and developing “Treasures of Islam: Art and Design in Islamic Manuscripts”. This trilingual multimedia CD-ROM (with interfaces in English, French and Arabic) contains rare or unique art work, together with a number of complete texts, selected from a collection of Islamic manuscripts held in the Department of Rare Books at McGill University. It combines text, still images, video sequences, music and voices. The paper deals with the following issues: Content selection (selection criteria, authentication); Digitization; Authoring system; Interface design; Information-seeking approaches; Multimedia integration; Multilingualism; Copyright; and Production and marketing. The paper also discusses CD-ROM versus the Web as a publication medium for this type of material. Manuscript Collections Manuscripts traditionally have posed major problems for libraries and archives charged with the twin but contradictory tasks of preservation and dissemination. Typically manuscripts are produced in limited numbers, and therefore any particular copy is likely to be rare, and perhaps even unique. This in turn may place a considerable financial value on the manuscript. Furthermore, manuscripts tend to exist in a more fragile state than, for example, published books. The manuscript may lack any kind of binding or cover, and indeed may comprise a single sheet of paper. If the manuscript is old it may also suffer from the various problems associated with paper and ink degeneration. It is therefore not surprising that many institutions place a very considerable onus on preservation. This can involve such measures as climate control and special storage containers. An additional major means to promote conservation, however, is simply to restrict access to the manuscript; each time a manuscript is handled it risks further physical degeneration. This can encourage policies ranging from refusal to lend the manuscript to other institutions, through restriction to use in a special room (where, for example, pens are banned), to the extreme where the manuscript is only available for display behind glass. Requests to photocopy or even photograph the manuscript may well be denied for fear that the manuscript will be damaged in the process. Such conservation practices are understandable with rare, fragile and valuable manuscripts. They do conflict, nevertheless, with the expectation that archives and especially libraries will actively promote the use of their collections. Some manuscripts may be important works of art in their own right, worthy of viewing much as might be a museum or gallery exhibit. But in many cases legitimate requests will be made to handle the manuscript, read it and copy it; refusal to accede to such requests for fear of damage will support the conservation role only at the expense of the dissemination role. Electronic replication of manuscripts offers one solution to this dilemma. Once the replication has been undertaken, limitless copies can be made with no degradation of quality and no need for further handling of the manuscript (in contrast to photocopying). When the importance of the manuscript lies in its beauty (for example, in its calligraphy or illustrations) the electronic version should be able to capture the full glory of the original. It may also be feasible and beneficial to enhance the original manuscript in the electronic version by, for example, scholarly commentary, indexing or translation into one or more languages. The copies then can easily be disseminated to remote users. Furthermore, the owners of the original may be able to generate income from sales of the electronic versions. It is not surprising, therefore, that a growing number of institutions should be turning to electronic storage and dissemination of rare manuscript collections. The Memory of the World project undertaken by UNESCO is an especially prominent example of a coordinated global attempt to safeguard and disseminate documentary treasures that are in danger (Courrier & Large, 1997). Such a solution, however, is not without its problems. This paper explores the issues involved in electronic replication of rare manuscripts, drawing especially upon the authors’ experience of designing and creating one such product. It also discusses the advantages and disadvantages of the two most popular electronic technologies for such work: CD-ROMs and the World Wide Web. Treasures of Islam The Islamic Studies Library, and the Rare and Special Collections Division of the Library at McGill University contain impressive holdings of manuscripts from the Islamic world. These include texts on jurisprudence, philosophy and theology in Arabic, Persian and Ottoman Turkish dating back as far as the thirteenth century. Many are in a fragile condition, inadequately documented and not easily available to the public. Islamic manuscripts are valued not only for their intellectual content but also for their physical beauty. Authors, calligraphers, painters and binders collaborate to produce works of art. As such, they provided a rich source from which to select the content of what would become the Treasures of Islam, a multimedia CD-ROM published in 1999 by McGill University. The Treasures of Islam CD-ROM comprises four sections: 40 examples of Islamic calligraphy in several different styles including Kufi, Thuluth and Naskh; 42 miniature paintings representing a variety of styles and sources; 16 richly decorated lacquer bindings; and eight complete manuscripts dating from the thirteenth to the nineteenth centuries. In addition, a short video clip and spoken summaries of each section are included. Three separate language versions of the interface are provided on the disc: Arabic, French and English. Not only the interface itself, but all supplementary texts (introductions to each section and captions for each image) and speech segments are available in the three languages. During the design and development of Treasures of Islam many decisions had to be taken concerning content, digitization, authorware, the interface, information seeking approaches, multimedia integration, multilingualism, copyright, and production and marketing. The solutions to these problems had to relate to the specific product under development, but the issues raised are generic and likely to be encountered by many multimedia developers. Content Although electronic versions of manuscripts can be used for preservation and dissemination, a considerable amount of time and effort must be expended in order to achieve this end. The content must justify the expense, and this is especially the case if it is intended to market commercially the product to cover costs or to generate income. Several selection criteria can be employed, including the rarity of the manuscript, its aesthetic interest, its financial value, and the level of demand for access to it. Decisions on content normally will be taken by subject experts rather than the technical designers. In the case of Treasures of Islam, for example, the Librarian of the Islamic Studies Library and the Curator of the Rare and Special Collections Division provided their expertise for content selection. Many of the manuscripts had not been well documented, and in order to present them on the CDROM in a meaningful fashion it was necessary for further research. Such an eventuality had not been anticipated by the developers, and indeed funds were unavailable within the digitization project for this kind of work. Fortunately, the Islamic Studies Librarian (Adam Gacek) generously was willing to undertake the necessary scholarship regardless. The lesson to be drawn here is that when planning such a project it should not be assumed that the manuscripts can be transferred to a new medium without provision for scholarly as well as technical assistance. Another consideration for content is the copyright status of the material selected for inclusion. The developers must be certain that they have the right to use the material without infringing copyright. Obviously this is especially important if it is intended to produce a commercial product for sale. Even if copyright is not a problem, the library or archive custodians may be reluctant to subject the manuscript to the rigors of scanning, a necessary step in digitization. Assurances may be demanded that every care will be taken with the manuscript, and it may be necessary to arrange that scanning be undertaken in the manuscript department, thereby obviating the need to remove it to the digitization project’s premises. The sound content of Treasures of Islam caused a number of problems. A decision was made to include background music, which naturally had to have links with the Islamic world. It proved difficult to find suitable music, even though only short segments of a few minutes’ duration were required for looping. Copyrighted music could not be used (unless copyright permission was obtained and any royalty fees accepted), and finally a local Islamic musician agreed to record for the CD-ROM (using the recording facilities of McGill’s Faculty of Music). Voice commentaries were required in the three languages of the CD-ROM, and after much experimentation with various voices and primitive amateur recording equipment, the Canadian Broadcasting Corporation (CBC) was hired to provide both professional readers and professional recording facilities. Digitization The manuscripts must be converted from paper to a digital version by scanning each page. Text can be stored as ASCII code very economically (a CD-ROM will hold around 250,000 manuscript pages). The situation is very different for still images, and especially for sound or video. The short high quality video clip of one minute on Treasures of Islam consumes 40 megabytes of disc space! Digitization is not only a slow process that must be undertaken carefully if handling rare and valuable manuscripts, but also one that calls for several decisions that will prove critical for the final product. In particular, a decision must be taken concerning the level of resolution to be employed. The higher the resolution, the better the quality of the resulting image. This is likely to be especially important if the manuscripts contain art work (as with Treasures of Islam). It should be noted, however, that image quality will also be affected by the kind of monitor and video cards to be used when viewing the product. Although high resolution will give a sharper image than low resolution, it can only be achieved at a price – the amount of disc storage space needed for the manuscripts. In other words, there is a trade off between resolution and space. An increase in resolution from 640 x 480 to 1024 x 768, for example, increases file size by more than 2.5 times. Although a CD-ROM will hold more than 650 megabytes, space can quickly be at a premium (this problem is alleviated with DVD technology that increases storage capacity seven fold). If high resolution is chosen then it may be necessary to compromise over the size of the manuscript collection or the way in which it is displayed (see below). Decisions over resolution are also influenced by the kind of hardware that users are likely to employ. If the CD-ROM is to be accessed only in-house then this information is available to the designers; otherwise they are compelled to take the best decision they can, balancing image quality against the size of the potential user community. For example, selection of low resolution (640 x 480 pixels) will ensure that the CD-ROM can be played on any of today’s computers. Although higher screen resolutions typically are now found (especially 800 x 600) many portable (laptop) computers still operate at lower resolution. It also ensures the best results regardless of how the resolution is currently set on the computer. On the other hand, if it is known that the CDROM will be used only on high-end desktop computers then a high resolution of 1024 x 768 will generate the sharpest images. It might seem a good idea to leave resolution choice with the user by providing more than one version on the disc. Unfortunately, this would require not only much more disc space but also the development of multiple interfaces, each one with its own screen design. Image quality is also affected by color depth. Sharper colors are obtained if the user’s computer is set on high color (16 bits or 24 bits per pixel) rather than eight bits. Authoring Software Software must be developed or more likely purchased to create the multimedia product. A number of suitable commercial packages are available, but the most popular (and the one used to develop Treasures of Islam) is Macromedia’s Authorware. This icon-based authoring program is used extensively in academic development environments, in large part because of its versatility in integrating multimedia components. Interface Design Interface design critically affects the usability of the program. Many authors have identified criteria for successful interface design. These include considerations such as the time it takes a user to learn the interface, the ease of retaining this knowledge once acquired, the time taken to issue an instruction and to receive a response, and so on. Interfaces should also be consistent, predictable and responsive to the user. (see Shneiderman, 1998, Galitz, 1997). It is both exciting and challenging to be given an opportunity to design an interface from scratch. The first step is to decide upon an overall design concept (or metaphor) for the interface. In the case of Treasures of Islam, a library building was initially selected as the metaphor; each of the four components comprising the content would be represented by a room within the library. Users could then “enter” a particular room and begin to select material from the shelves. This metaphor was later abandoned for the metaphor of a book (since the CD-ROM contained information about various aspects of manuscript books). Once the metaphor had been selected, the interface could take shape – the content was represented as four chapters, a title page and a contents page led into the contents themselves, the screens were designed as “pages”, the colors employed were chosen to give the impression of a manuscript book, and so on. Careful attention must be paid to screen layout. Cluttered screens should be avoided, and design motifs should be employed consistently throughout the interface. A lot of time can be spent on color and font selection. Color schemes are meant to enhance clarity in the interface rather than to construct a psychedelic display. Certain colors work better for backgrounds, others for lettering. Italic fonts are less effective on screens than on paper, and sharply defined fonts are to be preferred. Colors and fonts should be used sparingly and to draw user attention to specific activities on the screen. Above all, they should be used consistently throughout the entire interface. A critical element within the interface is the dialog transaction mode – the means by which the user can issue instructions to and receive responses from the computer. Traditionally this was accomplished via a command language, but more typical today are various kinds of menus, windows, buttons and icons, as well as various direct manipulation techniques. Information Seeking How does the user locate information on the CD-ROM? It would be possible to provide nothing more in the way of navigational tools than the capability to progress through the manuscripts sequentially one screen at a time. Such a minimalist approach would neither equal the possibilities offered by traditional print format (flicking through pages, opening at a random page, locating specific pages through a list of contents or a back-of-book index, etc.) nor exploit the many sophisticated possibilities available in a computerized version. Fundamentally, two navigational approaches are available (both can be employed, of course, with any one product): searching and browsing. Many designers will opt for a searching mechanism so as to allow specific screens to be retrieved quickly. This will involve the construction of an index through which individual words or phrases can be located in the text. Two approaches to index construction can be taken. The content can be examined by a human indexer to identify concepts that are considered potentially searchable by users. Alternatively, software can be used that will generate a keyword index of all the terms from the text, or perhaps only from certain segments of the text such as titles (typically excluding frequently employed “stop words” that do not convey subject information). The resulting index (or indexes) can then be presented as a browsable list from which terms can be highlighted by a user for selection. Alternatively, a search engine might be included that will allow users to input a term into a search box and then will search the index for occurrences of that term. The search engine may permit several terms to be combined using Boolean operators as well as offer features like truncation and phrase searching. Searching is effective to locate specific information to meet a well-defined user information need. In other cases, however, users have a less clearly formed idea of what they are looking for; they would prefer to browse through the information store, perhaps to get an overview of its contents, to see what they might discover by chance, or simply to read screens sequentially. Browsing can be facilitated by a number of devices. As mentioned under Interfaces (see above), menus, buttons and icons can be provided that allow users to move rapidly around the interface. Typically every screen will provide the means (a button or an icon) to move forward one screen, back one screen, to return to the opening screen or to exit the program altogether. In Treasures of Islam users can move quickly and easily to any particular screen within a chapter by clicking on the screen number pad displayed on the right-hand side of the screen (see Figure 1). An effective way to browse images (which are difficult to search except using an accompanying textual caption) is to display an entire set of images, reduced to thumbnail size, on one screen; the user can then click on any thumbnail image to select it in full size. Hypertext links offer a popular and highly flexible browsing mechanism, though in a large data store ancillary aids like a list of screens already visited, may be necessary to avoid disorientation.