On the Net, The Year Databases Moved to the Web

Greg R. Notess
Reference Librarian
Montana State University

ON THE NET

The Year Databases Moved to the Web

DATABASE, December 1998
Copyright © 1998 Online Inc.

Web access to these bibliographic database systems comes in many flavors.

Without a doubt, the World Wide Web is now the vehicle of choice for online databases and other collections of online information. Databases both large and small, created by established database producers as well as newcomers, have come to the Web. While some are available for free or supported by advertising, most of the traditional library databases remain commercial products.

While plenty of databases were available on the Web prior to 1998, this past year has seen further consolidation of Web-based search systems and greater sophistication from database vendors in their delivery of database information over the Web. Companies that did not have a Web interface for searching their databases have now added one. The companies that implemented Web versions before 1998 have mostly been busy improving their Web versions, moving customers to the Web, and adding new features.

BIBLIOGRAPHIC DATABASES

That staple of library resources, the bibliographic database, has been moving steadily toward Web accessibility. As libraries have gone from a paper subscription to a standalone CD-ROM, then to a networked CD-ROM, access via the Web is a natural progression. This past year has witnessed both multidatabase vendors and the individual database producers making more Web access opportunities available.

Database Vendor Approach

Advanced search features, limits, sorting, combining sets, and record display may be missing.

All the major search systems and CD-ROM vendors are making Web access a priority. SilverPlatter, Ovid, EBSCO, Information Access, UMI, and OCLC have several options for Web-based access to their databases. While Windows and DOS clients may still be available, with even a telnet version or two, the major development efforts are being focussed on the Web interface.

Web access to these bibliographic database systems comes in many flavors. The databases may still physically reside on a CD-ROM. The CD-ROM can be loaded locally. Another alternative for those with large hard disks--when the vendors allow it--is to copy the files from the CD-ROMs and onto local hard drives, then network the database from there.

Web access for locally-loaded databases brings with it all the requisite technical support pre-requisites that any other locally-loaded and networked database requires. The software needs to be installed and maintained. Local loading can allow more customization and easy changing of installation and setup options.

The other approach of the database vendors is to dispense with disseminating periodic copies of the databases and instead make it directly available via the Web. This remote hosting of databases has a number of advantages. Database updates can be much more frequent--as often as daily. The local institution does not need to have much technical expertise for support as the vendor maintains the software on the remote system. Large databases, especially those with substantial full-text components, are much easier to provide via remote hosting than trying to manage a local CD-ROM tower or changer. And the search software can be updated as soon as bug fixes or enhancements are implemented.

Unfortunately, the remote hosting creates other difficulties. Since the database is only in one location, if there is a system problem on that end, it will be unavailable to everyone. If the local Internet connection goes down, the database will be inaccessible. Response time and speed can be significantly slower than a locally-loaded database. There is also the lack of ownership of a copy of the physical database.

It can be more difficult to customize the interface of a remotely-hosted database with features, such as loading local holdings, choosing local preferences, or adding a company logo. However, remote hosting can manage all of those features. Vendors, including UMI, Information Access, and SilverPlatter, have all made some capability for local customization of remote databases possible with new capabilities introduced in 1998. Local periodical holdings, institutional logos, and local preferences, such as default fields to display, can be determined individually by each organization rather than making everyone accept the same package of features.

Database Producers on the Web

Beyond the multidatabase vendors' move toward embracing Web access, the second major approach to the Web for bibliographic databases finds the database producers bypassing the traditional vendors and creating their own Web-based products that they can market directly to their customers.

Pricing may be less expensive by going direct. However, the tradeoff usually is that the search interface is less sophisticated than one from SilverPlatter or the other multidatabase vendors. Advanced search features, limits, sorting, combining sets, and record display may be missing. On the other hand, some of these direct products do offer advanced searching as well. The local customization available from some of the multidatabase vendors is usually lacking from the Web products that come directly from the database producer.

Compendex Web (http://cpxweb.ei.org/), from Engineering Information, is one example of the direct approach and demonstrates one of the new possibilities for databases over the Web--end-user pricing.

NEW POSSIBILITIES

...the speed of response on a Web database is typically slower than the response from a standalone or even a locally-loaded, networked CD-ROM.

The ubiquity of the Web creates many new opportunities for disseminating and packaging of databases. From one single database, different versions can be easily created to appeal to different user groups and market segments. Many companies are exploring their ability to market database access directly to end-users for a fee, while at the same time offering institutional package deals.

Compendex Web sells the same product directly to individuals as well as to organizations. The Cumulative Index to Nursing and Allied Health Literature (CINAHL) provides another interesting example of this opportunity.

The CINAHL database is available from most of the major online and CD-ROM vendors. In addition, CINAHL also sells access to its own Web version of the database directly to both end-users and institutions. CINAHL Direct on the Web (http://www.cinahl.com/) is available in blocks of hours for individuals or by number of simultaneous users for institutions. Surprisingly, it has a relatively sophisticated search engine, with access to the CINAHL thesaurus, the ability to explode headings, limit to specific subheadings, and limit by publication type.

The version of LEXIS-NEXIS available as the LEXIS-NEXIS Academic Universe shows another way to take advantage of the opportunities offered by the Web. This product targets the academic market and only offers a subset of the full content available from the full set of LEXIS-NEXIS databases. Instead of selling this service directly to academic libraries, LEXIS-NEXIS has delegated development and marketing to another Reed-Elsevier subsidiary, Congressional Information Service, Inc. (http://www.cispubs.com/acaduniv/).

CIS has been very busy this past year signing up academic libraries that formerly could not afford to provide their students access to the full-text content available on LEXIS- NEXIS. Two aspects of CIS' approach stand out. First, most academic libraries that subscribed joined one of several very large groups for the advantageous consortia pricing. By doing this, CIS was able to sign up a large number of institutions with thousands, if not millions, of potential users. This leads directly to the second and more unusual aspect of Academic Universe: it will carry advertisements to help subsidize the reduced cost.

This will be an experiment to watch. The general Web search engines and sites like Yahoo! have amply demonstrated that advertising can go a long way on the Web toward paying for the development and maintenance of databases. However, the long-term success of that economic model still needs to be proven. It also has been difficult to get the advertising dollars spread out to sites that were not among the top dozen or so most popular sites on the Web.

With Academic Universe's focus on the higher education market, ads can be aimed directly at the interests of students and faculty in higher education. While CIS may not have been the first to think of the combination of subscription fees and ads (Electric Library and others are also using this approach), the range and scope of the Academic Universe database and its well-defined user group could lead the way to more advertising-subsidized database services. It could equally well prove that the model does not work.

Either way, the Web opens up numerous possibilities. If the present is any indicator, the way we deliver database access to our users in another decade may well look dramatically different from what we offered the past few years.

FULL-TEXT DATABASES

One of the major changes that is definitely occurring now is the shift from bibliographic to bibliographic plus full-text databases. From PubMed to OCLC's Electronic Collections Online to increases in the full-text portions of databases from Information Access, UMI, and EBSCO, access to the full articles online is the wave of the future. Unfortunately, the wave of the future is still breaking on a rocky shore.

There are many layers to the full-text puzzle. The vast majority of online full-text articles available through commercial databases continued to be in basic ASCII text during 1998. Some services, like UMI, offered a smaller number of titles available in full-image format. Full-image format typically means a scanned version of the article available on the Web in PDF format. Another approach is to add individual images to the ASCII or an HTML version of the text.

The Web makes delivery of all three options easy. At the same time, more publishers are actively putting up their own articles in full-text format. Access to these online articles is still by subscription, but some of the bibliographic databases are starting to link to these online versions.

Unfortunately, no one subscribes to all of the online journals. The ASCII versions often leave out important graphs or tables. The full-image version can be difficult to read. Some articles and sections are missed altogether. While 1998 has seen a greater number of periodicals available in full-text and some advances, there is still plenty of room for more growth and improvement.

WEB DATABASE SPEED

One of the most frequent complaints about the Web is how slow it can be. This complaint also holds true for commercial databases made available on the Web. The speed of any Web connection is determined by the combination of many factors: the connection from your machine to the local network, the speed from there out to the main Internet, the particular path the transmission must travel, the connection from the main Internet to the local network of the remote machine, and the processing power of that remote computer. The speed of database searching and manipulation takes a different face on the Web than we may be accustomed to with databases running on a local computer. On the one hand, the overall speed of access to information has been greatly increased, in many cases, by making the database available on the Web and thus accessible from home and office. Rather than having to travel to a physical library or other location housing a single computer, users can search a database directly from anywhere with a live Web connection. Countering this, the speed of response on a Web database is typically slower than the response from a standalone or even a locally-loaded, networked CD-ROM. Sometimes this is caused by a heavy load on the remote system, while at other times, it is heavy traffic on the Internet in general. Even when everything is running at optimum speeds on the Net, most Web-based databases seem to be a few seconds slower than a local counterpart. While the slight delay is manageable, the speed issue is likely to be one of continuing concern. As ever-greater effort is expended on improving Web database interfaces, we can expect a greater use of newer technologies. Ovid's Java client and SilverPlatter's WebSPIRS 4 rely heavily on Java and Javascript. XML may become very useful for bibliographic records. The potential that the Web has unleashed for providing access to commercial databases and links with related pages and full-text resources should make for a fascinating year ahead. Just stay alert as the familiar professional databases we use on a daily basis begin to change and adapt almost as quickly as the Web itself.

Communications to the author should be addressed to Greg R. Notess, Montana State University Libraries, Bozeman, MT 59717-0332; 406/994-6563; greg@notess.com ; http://www.notess.com.