Java Application Profile

Ivy-covered column Reader's Web
Museum Cafe
Managing 610,000 animal specimens with Java

Special to Intranet Design

The Museum of Vertebrate Zoology (MVZ) at U.C. Berkeley is a world-renowned research facility and the 4th-largest vertebrate zoology museum in the world. Founded in 1908, it houses a permanent historical record of mammals, birds, reptiles, amphibians and their ecology, evolution and geographic distribution. The MVZ's collections, in addition to providing important clues about historical changes in animal distributions, are the foundation for current programs in conservation, education and research.

Data about the Museum's 610,000 catalogued specimens is fundamental to both its own research and that of other research institutions around the world. In the spirit of academic collaboration befitting a university-affiliated museum, MVZ shares information about its specimens with other institutions free of charge.

Cartoon by J. Zakour and A. Noel Still, the manually intensive request process was anything but cost-free. Until recently the interested party submitted a written request for information, upon which the relevant data was researched and written up, resulting in a hardcopy report being generated and mailed to the requester.

Moreover, the data resided on a legacy flat file database that was more than twenty years old, running on a CMS operating system with a proprietary catalogue collections application named Taxir. Users of the system include curatorial assistants who perform data entry, and curatorial associates responsible for data integrity and augmentation. The associates also had the unenviable job of running detailed queries for researchers and members of the general public.

According to Barbara Stein, Curatorial Associate, MVZ, "The old database system was fast, but had limitations because there was no public interface, and all interaction with TAXIR was limited to two curatorial associates. Data integrity was also an issue and we had constant concerns of data corruption."

As these concerns and the drain on resources became increasingly burdensome the MVZ began looking for a way to streamline the request and delivery process.

Vertebrate evolution

In late 1997, armed with a grant from the National Science Foundation (NSF), the Museum IS department tested the waters for an Internet application development tool. After evaluating several, the team selected JADE from Vision Software Tools, Inc., an integrated Java development environment that includes a unique business logic server.

Vision persuaded the MVZ analysts that JADE, by leveraging integrated business rules and transparent database management classes, would enable them to build and deploy a new, Web-based collection management application much faster than other Java development tools.

The promises came true, according to John Wieczorek, MVZ project lead. "As a result of using Vision JADE we saw a 90% reduction in development time in the initial stages of the project. It would have taken about 3 months to complete the Java class packages on my own. With Vision JADE this was done in only 3 days.

"JADE allowed us to test ideas quickly and painlessly," adds Wieczorek. "This provided a sense of ... freedom that was non-existent during the 20-year reign of the legacy system."

Making information readily accessible to the public is another key project concern, readily met via the Web. Though the data is public, connections to the database server are carefully managed and controlled. Data has been migrated from the CMS system to a Sybase System 11 RDBMS running on SunOS 5.5.1.

MVZ is taking the opportunity during migration to enrich its collection schema, expanding from 91 fields in the flat file to 106 tables having a total of 516 fields in the RDBMS. This factor of 5 increase positions MVZ to give meaning to data once "hidden in remarks fields," as well as data for auxiliary collections, which have not been digitally managed to date.

Reusable components

At the MVZ, application needs vary greatly, but all must adhere to certain common interface guidelines to maintain consistency and reduce the burden on both users and developers. MVZ benefited in this regard because Vision JADE supports reusable objects with embedded business logic.

Other Java development tools require laborious hand-coding to implement logic in event handlers and component methods. In JADE, much of this functionality is built into a Business Logic Server (BLS) that can be programmed with natural language business rules rather than code.

JADE's unique foundation in objects with an emphasis on business logic makes it highly suitable to another aspect of the MVZ effort: serving as a pilot for the Museum Informatics Project (MIP) at U.C. Berkeley. A computer resource center for 20 museums on campus, MIP is evaluating various software solutions in an effort to create a unified approach to meeting the various museums' information needs.

The MVZ is central to MIP's concerns as it represents a large, internationally respected research facility with in-house expertise and a clear model of museum data needs. U.C. Berkeley is aiding in MVZ's efforts and applying the methods to other museums on campus as they begin to automate their data processes. While these museums are not yet committed to using Vision JADE, MVZ's success coupled with its growing library of reusable classes supports adoption of JADE as a MIP standard.

With a supplemental grant, the MVZ is prototyping a system for referencing field notes and photographs associated with specimen records. The scope of the project is in the area and time period written about in "Animal Life in the Yosemite," by Joseph Grinnell, founding Director of the MVZ. Vision JADE will be the development platform.

Nor does the theme of managing museum collections and expediting information exchange stop at U.C. Berkeley's campus boundaries. The National Science Foundation and Natural History Museums around the world are interested to see the fruits of the MVZ project in order to leverage the development efforts.

In particular, The Natural History Museum at the University of Michigan, The Burke Museum at the University of Washington, The University of Alaska Museum, and The Smithsonian Institution are anxiously awaiting MVZ's results in the hope of leveraging its data model - a database schema fostering component model reuse and reapplication - for their own museum collections.

In effect, a broad community is watching the implications of what MVZ is doing and the implementation has far-reaching effects. This project is important, not only to MVZ, but also to the natural history community as a whole.