Global Memory Net Offers New Innovative Access to Tsurumi s Old Japanese Waka Poems and Tales, and Maps

Transcription

1 Global Memory Net Offers New Innovative Access to Tsurumi s Old Japanese Waka Poems and Tales, and Maps Takashi Nagatsuka 1 and Ching-chih Chen 2 1 Tsurumi University, Dept. of Library, Archival and Information Studies Tsurumi 2-1-3, Tsurumi-ku, Yokohama , Japan nagatsuka-t@tsurumi-u.ac.jp 2 Simmons College, Graduate School of Library and Information Science Boston, MA 02115, USA chen@simmons.edu Abstract. This paper describes how Global Memory Net (GMNet) has been able to provide new kind of innovative access to the invaluable content of the classical Japanese ancient poems and maps at Tsurumi University that was not available for public access before. The collaboration began with the development of a prototype collection, based on images included in two publications of the Tsurumi University Library - the Eighty Selections of Waka Poems and Tales from the Classical Japanese Literature and the Japanese Maps in the Old Age. As the project developed, coinciding with the technology development of GMNet in bilingual retrieval as well as with sound presentations, the inclusion of sound files for each of the Waka selection was considered a very desirable feature since Waka poems are generally only readable by very small number of specialists. The paper will present a bird s eye view of how Tsurumi s rare collection is organized, presented with much enhanced access in GMNet system. Through this project, the Tsurumi team has gained considerable important experiences. The overall process for them was very time consuming even though the technology of GMNet was already in place. These valuable experiences will be discussed and shared. 1 Introduction This paper is to describe how by taking part in the Global Memory Net (GMNet), one can gain new innovative access to the invaluable content of the rare classical Japanese ancient Waka poems and maps at Tsurumi University that was not available for public access before. The possibility to participate at the GMNet became a reality after Ching-chih Chen s visit to the Tsurumi University as the first invited speaker at the University s initial Digital Library Symposium in late December 2004 [1]. At that symposium, Chen stated that GMNet aims to be an effective gateway or digital portal to the world culture and heritage. It can also be viewed as a comprehensive multimedia digital library which can offer the global culture, history, and heritage instantly to the information seekers [2]. She indicated that Japanese contents are not yet sufficiently accessible by people in the world due to language and other factors, and that GMNet

2 sible by people in the world due to language and other factors, and that GMNet can provides a platform to make it possible for global sharing and distribution of some selective Japanese resources. She also pointed out GMNet s great potential for real collaboration with different countries in the world not only in the content areas, but also the needs for involving subject specialists in those countries in order for GMNet to provide more in-depth knowledge on the significant cultural and historical memories of those countries [1]. With this impetus, Tsurumi group led by T. Nagatsuka welcome the opportunity to join GMNet by benefiting from its cutting-edge technology to develop a prototype for resource sharing and for promoting universal access. 2 Development of a Prototype Tsurumi Collection Tsurumi s collaboration with GMNet began in early 2005 with the development of a prototype collection related to the rare Waka poems and tales, as well as ancient maps on Japan. The value of ancient maps on Japan is clear, however, why Waka poems and tales? Waka poems and tales are rare and important classical Japanese literature, thus an invaluable part of Japanese cultural heritage. Hundreds of original texts, translations, studies, and electronic texts of classical Japanese literature are available on the Web, and many are in Japanese only. As to the Waka literature, there are several useful Web sites as well, such as the Japan 2001 Waka Website by the University of Sheffield which translates some Japanese texts of the imperial waka anthologies to English ( and the site of the Japanese Text Initiative of the University of Virginia Library s Electronic Text Center which displays classical Japanese literature including Waka poems in Japanese characters ( Yet, none has introduced the original look of the handwritings of the poems, presented sound files of the contemporary Japanese reading, and permitted bilingual retrieval capabilities. These are the features of our prototype in GMNet. The Tsurumi collection includes images chosen to be included in two publications of the Tsurumi University Library -- the Eighty Selections of Waka Poems and Tales from the Classical Japanese Literature and the Japanese Maps in the Old Age. The first book, Eighty Selections of Waka Poems and Tales from the Classical Japanese Literature, includes images taken from detached segments of Waka poems and Tales selected from royal anthologies, poetic anthologies, prose narratives, tales of war, panels of screens, and Japanese backgammon and playing cards. These were handwritten or printed from the end of 11 th to 19 th centuries. Each image represents a part or the whole writings, and Japanese description of this portion is included in the back part of the book, which form a good base of the metadata of the image. The second book, the Japanese Maps in the Old Age, is composed of thirty-seven ancient maps on Japan published in Europe and three in Japan. Each map image comes with descriptive title and associated bibliographic information in both English and Japanese. The reasons for selecting these two books were:

3 The cultural contents of the two books were considered to be the best of the Tsurumi University Library s rare collection. They are unique, rare, and appropriate for meeting the criteria for inclusion in GMNet. The prototype development will need the involvement of the subject specialists in Japan. The contents of the books are made up of images with Japanese descriptions. Without English descriptions, it will be of lesser use to non-japanese people. In handling both English and Japanese, this offers new challenges to GMNet for demonstrating its bilingual capabilities. The rare Waka poems and tales are not readable by most Japanese, and thus, the need for currently reading by experts in the field is great, and this allows the GMNet to explore the retrieval of sounds in addition to images. 3 Developmental Process As the developmental work starts, all images were scanned by Chen s research group at Simmons College in the US. Also, the baseline database information in English, when possible, was completed by Chen s research staff as well using the GMNet s existing metadata structure. 3.1 Metadata and Descriptive Information After the creation of the initial baseline database, it was found that several metadata elements will need to have the Japanese translated by the Tsurumi University staff. Specifically, the descriptive information on each image of the handwriting page of the Waka poem or tale has to be prepared in English for the Description field of the metadata. This has proven to be a very difficult task, because most Japanese descriptions could not be translated to English directly, otherwise they would not be understandable by readers without possessing the basic knowledge of the classical Japanese literature. Thus, the Tsurumi staff had to create new modified descriptive information in English for every image. To complete this job, it was necessary to consult and discuss with subject specialists at the Tsurumi University for each description. Although the descriptive information for each record is short, yet it is a very labor-intensive process. It is also a difficult one because there is an obvious cultural difference between the staff in information science and that in classical Japanese literature studies. Each has his/her own emphasis. For example, the kind of English description needed for GMNet does not correspondent to that considered to be important by the classical Japanese literature specialists at Tsurumi University. To prepare for the use of bilingual retrieval functions, the metadata in Japanese language is also prepared for the 80 Waka Poems and Tales. Except the description field, all other metadata fields in Japanese are created using the same content as those in the English version using the GMNet s existing database structure. The Japanese description follows the same contents as published in the book.

4 3.2 Tsurumi Collection in GMNet In this section, we shall provide a bird s eye view on how one can access Tsurumi Collection via the web-base GMNet. Figure 1 is the introductory page of the Tsurumi Collection. As indicated there, The process in developing this digital collection is a perfect example of both difficulties -- distance, cultural differences between the subject and the technology specialists, language difficulties, ability to interpret ancient contents, etc. -- and exciting results of international collaboration. In meeting the challenges, GMNet has extended its image retrieval capabilities to include the retrieval of sound files, and also begun to experiment bilingual retrieval. To explore bilingual retrieval capabilities, we have included the metadata in both English and Japanese in UNICODE. The Japanese version is produced by the Tsurumi team using our metadata structure. Users can thus search in both languages if they have the Japanese input capabilities on the browser. Figure 1. Introductory page of the Tsurumi Collection in GMNet When one enters the Tsurumi image collection, the screen as shown in Figure 2 will show up. Where a user can decide to find what they want by searching in any of the metadata field or all. It can also search by using the cutting-edge content-based image retrieval technique, SIMPLIcity of the Penn State University. Here, one can

5 browse images of the collection, or can randomly access the image collection. When the images are displayed, simple titles of these are also shown (Figure 2). Figure 2. Images of Tsurumi s Waka poems with English titles When a specific image of the Waka poem is chosen, additional descriptive information of that poem can further be obtained by clicking on the image, as shown in the left panel of Figure 3. In addition, if one would like to see the current Japanese translation of the ancient writing, one can go the pdf file directly (Fig. 4). When larger is clicked, the image can be zoomed and enlarged several times to enable the seeker to view the details of the handwriting of the Waka poem. To protect the ownership of the image, dynamically generated digital watermark can be seen in the lower right corner of the enlarged image (Figure 5). Fig. 3. Metadata with detailed descriptive information

6 Fig. 4. The pdf file of the Japanese translation and transcription of the Waka poem Fig. 5. The chosen image can be enlarged several times depending on the image resolution. Note the dynamically generated digital watermark in the lower right corner.

7 3.3 Bilingual Retrieval As the project developed, coinciding with the technology development of GMNet in bilingual retrieval, GMNet naturally explore the retrieval of the Tsurumi collection in both English and Japanese. If a user can input the Japanese characters from his/her PC with a bilingual wordprocessing software, it will be able to access to the Tsurumi Collection by Japanese. After the user retrieved in English and received the output in English, he/she would like to read more detailed description in Japanese. In this case, he/she can retrieve the descriptive information in Japanese. Figure 6 shows the database information in Japanese on the front screen, while the English is on the back. Figure 6. The bilingual retrieval of images in the Tsurumi Collection in GMNet. 3.4 Sound Presentations As GMNet is developong bilingual retrieval capabilities, it is also exploring the use of sound presentations. It was considered that the inclusion of sound files for the Waka selection is truly a desirable feature since few people, even Japanese, are able to read the old and ancient Japanese poem. Thus, the sound files for most writings were created by recording the readings digitally and then converted to Winamp (Windows Audio MPEG Player) media files for the web presentation. As shown in Figure 7, when a user retrieved the Japanese transcription in pdf file shown in the upper right panel, the sound file narrating that transcript can also be retrieved as shown in the lower right panel.

8 4 Conclusions Fig. 7. The sound presentation of Tsurumi Collection in GMNet The description above covers only the tips of an iceberg of the system capabilities of GMNet as applied to the Tsurumi collection. Yet, it is sufficient to show that, as the results of cooperating with GMNet, for the first time, the old Waka poems and tales, as well as ancient maps of the Tsurumi University not only can be shared with those scholars and learners who are interested in classical Japanese culture and heritage, but also greatly enhanced their capabilities in obtained needed multimedia information quickly and effectively in a way not possible before. With the cutting-edge technology of GMNet, Tsurumi s collection is now available for universal access with many attractive features as an exciting part of the global gateway. However, the seemingly small number of records with only the provision of limited metadata information have consumed lot more time and effort than originally anticipated by the Tsurumi staff, even though the technology-related work has been handled and completed by Chen s GMNet group. Through this project, we have gained considerable important experiences. The overall process was very time consuming, especially there was much difficulty in the process of preparing the raw English descriptive data. One of the major reasons is the information science specialist s lack the subject knowledge to the classical Japanese literature. This kind of difficulty will be experienced by any other similar types of

9 projects. The other reason is the subject expert s understanding and knowledge on the purpose as well as use of digital library applications. Basically two group s purpose and focus are not similar. What we should do next? will be a significant question for us to consider. There are numerous digital library projects for preserving cultural heritage in Japan, but great majority of them are mainly targeting to the Japanese audience or the Japanese information seeker. Clearly to create the English access capability is essential for international sharing and information distribution, but it is also of great difficulty at the moment. Tsurumi s participation in GMNet not only confirms this problem, but also suggests that there is a great and urgent need to explore possibilities in meeting these challenges. This pilot project is an important step for Tsurumi in international collaboration for introducing the Japanese cultural heritage, especially the classical Japanese Literature, to the world. We now know that we have much work ahead of us if we are serious in pushing our digital library initiatives. Aknowledgements We are grateful to Tsurumi University s Faculty - Takashi Horikawa, Fumito Ikura, Kyoko Yamakawa - for their effort and advice in our creation of the English descriptions of the Waka poems and teles. Special thanks goes to Kyoko Yamakawa for narrating the sound files. We greatly appreciate the efforts of several research staff of the Chen s GMNet group at Simmons College. Specifically, Annie Cain did all images scanning, Peishan Tsai completed the much needed metadata work. Shengqiang Zhang of Sichuwan University in Chengdu, China and visiting researcher with Chen s group developed the bilingual capabilities in addition to the current PHP-based system under Chen s overall direction. References 1. Chen, Ching-chih, Global Memory Net offers the world instantly: Potentials for universal access to invaluable Japanese contents, Journal of Information Processing and Management (in Japanese) 47 (11): (2005). Translated to Japanese by Takashi Nagatsuka. 2. Chen, Ching-chih, Global Memory Net: New Collaboration, New Activities and new Potentials, in Proceedings of ICADL 2004, Shanghai, December 14-16, Edited by Z. Chen et al. pp