Publishing Textual Databases on the Web
Jacso, Peter, Information Today
Peter Jacso is associate professor of library and information science at the department of information and computer sciences at the University of Hawaii. He won the 1998 Louis Shores--Oryx Press Award from ALA 's Reference and User Services Association for his discerning database reviews. His e-mail address is firstname.lastname@example.org.
Part 2 examines HTML options for moving from the desktop to the Web
Publishing a database using some of the advanced features of HTML 3.2 is the least expensive (although not the simplest) approach to make an existing collection of bibliographic, full-text, or directory data accessible on the Web. Often, it is also the only possibility for those whose Internet service provider does not allow the installation of Common Gateway Interface (CGI) scripts on the server. The reason for the prohibition is that CGI scripts can do a lot of harm on a server by deleting files and crashing the system. The pure HTML options have certain limitations, but they provide inexpensive tools for individuals and organizations with tight software budgets, such as libraries and information centers. The choices among the pure HTML, Java, and CGI or DLL (Dynamic Link Library) options are not exclusive. Neither are they exhaustive.
For a masterpiece example of highly functional and intuitive design that will please all the users all the time, see the Scholarly Electronic Publishing Bibliography (http://info.lib.uk.edu/sepb/sepb.html) of Charles Bailey, who pioneered the use of the Web in libraries and by librarians sooner than many of us could spell HTML correctly. This excellent bibliography is matched with equally excellent options that allow users to print in Microsoft Word or Adobe PDF format, search via Boolean operators, or navigate through well-structured HTML pages composed of and linked by major topical categories (Economic Issues, Legal Issues, Case Studies, etc.). I will illustrate the HTML option here using a simple Web prototype of an abstracting-and-indexing database that was created originally for local use and CD-ROM distribution.
Repurposing the Original Database
The abstracting/indexing database in its desktop version filled a niche by providing access to the most recent 12 volumes of the journal Social Process in Hawaii (SPIH) by authors, author affiliation, descriptors, title and abstract words, and a combination of these. Such access to this valuable local source was not available at the conception of the project in any form. The database was created by students in my Abstracting/Indexing, Database Design, and Database Publishing courses. It was enhanced by adding searchable full-text records and page images for some of the articles. Currently, I am working on a simplified HTML-only version of it to demonstrate that a well-organized bibliography in pure HTML (although using frames) can be a useful and valid alternative to the fully-searchable desktop and CD-ROM database versions.
If you have created a bibliography with any bibliography formatting or database software, you are able to print the bibliography and the indexes to a file. This file in turn can then be fed to a decent word processor that is endowed with HTML output capabilities. Up to this point the process is a no-brainer. However, a linear bibliography of several hundred screens would not be efficient to consult. It must be organized and broken into smaller units, such as subject bibliography (subj.htm) and author bibliography (auth.htm), and possibly into further subsets by the letter of the alphabet (subjA.htm, subjB.htm,). These subsets also can be easily created using the method described above.
Choosing Primary Access Points
The HTML documents can be linked through a main page to allow users to consult first the HTML list of subject terms (descriptors) or authors. These in turn would display an index of the articles that have the selected subject term or author, respectively. …