AsianDOC Electronic Newsletter 1:1(March 1998)
Michael West Wiemer
Stanford University
Kathryn Akiko Kada
University of California, Berkeley
Richard B. Dasher, Ph.D.
Stanford University
Currently, the J Guide presents four directories of resource tools (News, Search Engines, Indices & Guides, and Reference), twelve topic directories (discussed below), and one subdirectory in an area of special focus (Nanotechnology). Current usage is approximately 24,000 hits per month.3 Links for email feedback, which are available on every page, generate questions and requests for specific information, as well as comments, every day from all over the world.
Since the inception of the X-Guide in 1994, the J Guide has been and continues to be developed primarily by graduate and undergraduate students under the general supervision of Stanford faculty. Therefore, turnover in personnel associated with the project has been relatively rapid. This has posed considerable challenges to the preservation of consistency in information classification and website maintenance. Moreover, the combination of (a) the explosive growth of Internet-based Japan information, (b) the appearance of new Internet resources such as Yahoo! and Yahoo! Japan, (c) new browser and database management technologies, (d) a small budget, and (e) the need to coordinate the J Guide project within the broader research and educational objectives of the university have led the project to develop a unique approach, balancing limited comprehensiveness with greater depth in selected topic areas. Not surprisingly, near-term project goals include increased automation and better resource tools for future developers.
Following a description of the background of the project, the sections below describe the approaches we have taken to meet the above challenges. They will also show the insight the USJTMC has gained in the demand for Japan information, and into how we can tailor our services to the ever growing and changing audience on the Internet.
At that time, only about half of the Japan information sources listed in the X-Guide had publicly accessible Internet websites; other entries were plain text lists of organization names and contact information. Since then, however, the number of websites in Japanese companies, government, nonprofit organizations, as well as those of private interest have continued to grow at a very rapid pace. Accordingly, one major task in updating the X-Guide in its early stages was the replacement of plain text lists with hyperlinks, as websites from the various information sources came on-line.
From the beginning, USJTMC developers actively sought out information sources, both plain text and websites, for inclusion in the X-Guide. Following public release of the X-Guide, users likewise began to play an active role by recommending websites for inclusion in the directory. Their recommendations indicated that there was major demand for a broader focus to the directory. Similarly, USJTMC developers were also led into new areas as they researched for answers to user inquiries for information about Japan. One major area of expansion has been in "Culture" content. The X-Guide had one topic for "Culture and History"; at present, there is so much content under that classification that it has been distributed across three different topic areas: "Language & Culture," "Art & Entertainment," and "History and Geography." Other changes in the topical arrangement of X-Guide into J Guide are discussed below.
The primary reason for renaming X-Guide to J Guide was to show that the initial experimental project had developed into a more complete, ongoing website. Also, with a name like X-Guide, it was not surprising that the website received some misguided hits by those expecting to find "X-rated" information. To avoid this misconception and with the sense that it was no longer necessary to term the directory "experimental", we officially changed the name of the directory in late 1996. At this time, the project also embarked on a major redesign and on critical assessment of the topical categories. One of the major design choices proved to be whether or not to utilize "frames" technology.
The USJTMC development team experimented with a frames version of the J Guide main directory for approximately seven months. Frames are definitely a convenient way to display an index of main topics for easy access from anywhere within a given site. The main advantage to a frames design was that it kept the topic categories in view, along with each page of links. Also the topic categories remained in view even when the user had accessed a third-party website from the J Guide directory. The pages of links and third-party websites appeared in an inner "content" frame, while the remainder of the frames served as our navigation aide. This layout seemed to be the best way to maximize mobility within our collection by providing users with an easy, visible route by which to return to the J Guide.
Throughout the experiment, we made both a frames version and a non-frames version of the J Guide available. The frames version, however, received some mixed feedback from our users, and we ultimately decided to discontinue it. One problem was that, in the technology at that time, the browser header continued to display the URL of the J Guide, regardless of whether or not a third-party link was being displayed within the inner content frame. We received some user comments that indicated that the users were unclear about the ownership of the information displayed. While we make our best efforts to link only to accurate, useful websites, we are of course in no position to guarantee the timeliness or accuracy of all links in our directory. Unfortunately, the frames design blurred the lines of ownership and responsibility. A second problem with the technology was that addresses of pages in inner frames were not easily bookmarked by users.
The X-Guide topics reflect the pioneering nature of the site at the time. Some topics contained considerable overlap, e.g. "List of Japanese Corporations" and "Business, Economic, and Financial Information". Others were really more about U.S. organizations than directly about Japan information. As websites in Japan and Japan-related websites in the U.S. proliferated, it became important to refine the clarity of focus of each category, systematize the topic classification according to some standard, and make it clear what types of information we would and would not include in the directory.
With regard to selecting information not to include, some categories were removed because the information became available elsewhere. For example, "About the US-Japan JITMT Program" became unnecessary when an official JITMT homepage was launched on another website. "About the Stanford US-Japan Technology Management Center" likewise became available in another portion of the USJTMC homepage upon its redesign.
J Guide topics at present consist of the following:
As shown above, a major addition to this website since the X-Guide days has been the inclusion of the tools directories: News, Indices & Guides, Search Engines, and Reference. While now a common feature of almost any Internet directory, the content for these categories basically did not exist at the time of the beginning of the project. A commonly asked question is the differentiation between the categories of "Indices & Guides" versus the "Search Engines." There is a definite overlap between the two categories. However, the distinction is made to emphasize whether the tools are of directory nature (incorporating a breakdown by subject area) or the keyword-search type engines, respectively.
With regard to the main topics, the project has seen a number of rearrangements, additions, and deletions of categories over time; these sometimes reflected the different opinions and interests of different developers as project personnel changed over time. The current category topics, however, are not random divisions. In order to systematize the main topics, the USJTMC development team undertook a survey of different library classification systems with the purpose of considering their application to the J Guide content. The two that were looked at most closely are the Library of Congress and Dewey Decimal systems.
We decided, however, that it would not be practical to incorporate fully either library classification system in the J Guide. The two library systems are primarily for book classification, rather than for the constantly changing information found on the Internet. Moreover, the domain of J Guide information is all related to Japan. Because of the nature and number of links that need to be included in some categories of J Guide information, it made more sense to maintain separate listings for them. Thus, on our site "Computer & Internet" continues to be a distinct heading on a level with "Science & Technology." Nevertheless, the process of studying the two library classification systems has given us a better view of standard subject headings, and it has led to a deeper understanding of the structure and requirements of our directory.
When the X-Guide was first released, the most common web- browser was NCSA Mosaic, before a multilingual version was available. As multilingual computing capability has spread, and as the number of Japanese language websites has increased dramatically, we have begun to incorporate links to Japanese language sites as well as English language ones. This practice is discussed further below.
Along with the fact that the number of websites has exploded, there has been increased interest in different areas by our users, as seen from email comments and inquiries directed to this site. It soon became clear to our small development team that it would be impossible to attempt to keep up with adding all new websites pertaining to Japan. The following section addresses our approach to solving this dilemma.
Therefore, the USJTMC has opted instead to place effort into adding value by carefully choosing a subset of links to list for any given area, by tracking access statistics carefully, by focusing on areas of special focus for new links, and by rationalizing some of our other web-related activities under the J Guide. In addition, we have come to rely greatly on our users as an important resource: users provide us with the information for many new links and needed updates.
A further service that has become possible and also needed in light of the rapid spread of Japanese language websites is the linking to (and differentiation) of the Japanese language and English language versions of a website. This service is of value not only to educated native readers of Japanese, but also to people who are studying the language or who need to find information, such as a recent news release from a Japanese company. With the proliferation of websites in Japan, it has become a fact that many Japanese language websites are more up-to-date and contain more information than do their English language counterparts, concerning information related to Japan. The link to the Asahi Shimbun with both Japanese and English links (below) illustrates our approach to solve this problem.
Yet another service is to provide comments to characterize the degree of activity on a particular website. Certain kinds of information require comments as to a link's activity. For example, in tracking a developing technology, such as "Nanotechnology Research and Development in Japan," regular users of J Guide want to know how often they should revisit a particular site to which we link. Therefore, J Guide is now incorporating brief indications of when a page was last updated, how often it is updated, which sites are new, and in some instances the contents of a page.
This page is currently the leading set of links in the world for Japanese Nanotechnology R&D, with almost 100 links to pages that directly deal with institutions conducting nanotechnology research in Japan or analysis of the Japanese research efforts. Our efforts to develop such areas of specific focus are an attempt to fill gaps in the worldwide research community's relative lack of knowledge concerning Japan activities. Other major directories in the U.S. and Europe of nanotechnology websites still list only one or two links to specific Japanese research projects.
The success of the Nanotechnology directory since its debut is shown in the graph below. The USJTMC has not actively advertised this new page in search engines or in other directories, besides sending out an email to members of a nanotechnology discussion list that the USJTMC hosts. Nevertheless, as shown above, the number of hits to this page are growing rapidly, relative to other pages on high technology that the USJTMC maintains.
The red line in the graph represents the rate at which hits are accumulating for the nanotechnology page, which was started on January 23, 1998. The first rapid increase in hits (on January 28) may correspond to the response to the announcement of this new page by the email distribution list, or it may be due to search engines having indexed the page--and users accessing through those search engines. Since then the number of hits has been increasing at approximately 15 hits per day. With the release of the nanotechnology page the other pages have also experienced an increase in their hit rate, illustrating the point that users are enticed by neighboring links. Most significant is the fact that the introduction of the nanotechnology page has caused a noticeable increase in the hit rate for the main page of the "Science & Technology" category. This is most likely due to the fact that the nanotechnology page is a subtopic on the S&T page. Note however that the hit rate for the main S&T page has increased in excess of just the addition of the nanotechnology page hit rate! Perhaps this shows that users are also easily side tracked by neighboring links, or realize that the site they came to does not appear to contain the information they were seeking.
Current J Guide statistics reveal an interesting dynamic for this website: that the popular topics change monthly, and the number of hits each topic receives varies greatly. In order to present the data in a comprehensible form, only the most current data from December 1997 to February 1998 is included. Yet, both the "Computer & Internet" and "Indices & Guides" categories are not included in the top 10 most frequently accessed topics for all three months. Thus following user demand for a certain kind of information seems an impossibility based on the statistics gathered. This data may also be a reflection of the rapidly growing and changing population of users on the Internet.
The "News" category has the highest number of hits on average for the three months in question. Users undoubtedly have a desired to gather the latest information about Japan from the well known media sources. Perhaps just an artifact, but the "Life" category has received the highest number of hits this February for any category in the last three months (and the same in January). Could it be that users are planning their spring and summer trips to Japan? Another interesting note is that three of the four Tools directories are included in the most frequently accessed sites. Information gathering tools are clearly a necessity for navigating today's web.
BaddressBot was created by Norman Wayne Karo, with input from Michael Wiemer and other USJTMC staff, as an educational exercise in Visual Basic 5.0 web development techniques, and to meet a practical need of the USJTMC: that of determining the dead links in the J Guide. It is a web robot based on "The Web Robot" and "The Parser," and modified to meet the specific requirements of processing a website similar to the J Guide in a robust and efficient manner.6 A sample screen of the BaddressBot appears below.
BaddressBot's simple windows interface allows any user to quickly check any website and determine the status of the links contained therein. BaddressBot follows the standards for robots exclusion, so it is "polite" when visiting other sites and will not repeatedly attempt to access sites where it is not wanted. Its basic functionality is that it processes a given html page for all of the links on that page to image, text, and html files. After extracting these links it checks the status of each page, reporting back if the file no longer exists, the server is not accessible at the moment, the file has a notice that it has moved, or that the link is in working order. An indented list of the addresses visited can then be viewed and saved. This list is of use to the web administrator for maintaining the webpages in question.
A second opportunity for rationalizing our efforts has come through negotiations to combine our efforts in regard to two separate Japan- related directories. The early success of the X-Guide led the USJTMC to be entrusted with the World Wide Web Virtual Library entry for "Japan". The WWW Virtual Library has been an ongoing project of the USJTMC for several years; its contents are similar to that of the J Guide. So, to consolidate efforts and provide a unified source of Japan information resources, the USJTMC has negotiated the incorporation of the WWW Virtual Library entry for "Japan" under the banner of the J Guide. The more comprehensive J Guide will continue to deliver ever more useful content to the WWW Virtual Library, and the name and logo of the WWW Virtual Library will provide further recognition of the quality and value of the J Guide. Content differing from the J Guide has been incorporated therein.
But in the world of electronic information, one of the most important services that can be provided is that of increasing contacts and human relationships. The USJTMC plans to try to do just that through interactive features which will be added to the J Guide. By connecting experts in narrowly defined fields, and providing their knowledge to others, it may be possible to help users keep current with the hyper-abundance of information being presented on the Internet--and allow the access of information not presented electronically. Also categorizing and maintaining hundreds of links requires effective use of technology, the benefits that come with careful automation.
However even though technology to display Japanese text is advancing, most people lack the necessary language skills to read that text. To retrieve much of the detailed information on Japan that is in Japanese, requires a high degree of proficiency. Several tutorials have appeared on the web that teach fundamental Japanese or even specialize in teaching proverbs, kanji, or similar material. However there is a lack of web sites that provide tutorials on advanced Japanese and technical topics that are often necessary for successful information retrieval.
Moreover, as machine translation software and services continue to improve, other skills such as the ability to locate quickly a desired Japanese language webpage and accurately grasp its overall relevance to the reader's goals become important as well. It is the opinion of the authors that integrated tutorials would be valuable in leading online students through the fundamentals to acquire such skills for advanced Japanese, with concentration on reading as it applies to the Internet. Likewise, software tools that provide electronic dictionary and thesaurus functionality are also of great importance. Resources that help bridge the language barrier will provide individuals with the ability to continue the pursuit of Japan information.
Providing this kind of information online in an internal website is ideal in that the information is accessible by everyone in the group, and the information is centralized in an easy to find area.
Therefore, as a major improvement to the quality and regularity of the J Guide, the USJTMC anticipates moving to a database platform as the basis of the J Guide. The initial system which will contain the J Guide is a stand alone database, from which static html pages will be generated, as opposed to a dynamic interface using a cgi or similar script which interfaces with the database. The upgrade to a database from which we can generate static html is the logical choice, considering the number of links currently in the J Guide. This step will also serve to test the database structure before moving on to a system that can handle even more links and provide dynamic html pages.
Advantages of the database system in question are that updates to html pages are easier, consisting of entering data into the database and selecting an option in a program that builds the html page from the data. Others not familiar with html or standards for the pages in the J Guide will still be able to generate pages with minimal instruction. Additionally searching for links is possible by any field in the database, whereas currently the web administrator has to visually inspect several pages until the link in question is found. Perhaps the most advantageous point is that links will be stored in a central place that permits easy checking by automatic scripts without the need to parse html files.
Current research is limited to a database model that should be able to contain the current J Guide and expand the information provided.
The USJTMC will continue to provide a quality source of Japan information resources with the J Guide; we plan to execute the expansions already mentioned, in the near future. Therefore periodically, the USJTMC will submit updates concerning the development of the J Guide and the knowledge we have gained in this process.
URL http://asiandoc.lib.ohio-state.edu/v1n1/dbs/jguide.html
|