ELSA and MORE: A Library and an Environment for the Web

Submitted to: The 13th Annual National Conference on Ada Technology

Keywords:

Repository Management

Software Reuse

World Wide Web

Authors:

David Eichmann**, eichmann@rbse.jsc.nasa.gov

Margaretha Price*, mprice@rbse.mountain.net

Robert Terry*, rhterry@rbse.mountain.net

Lou Welton*, lwelton@rbse.mountain.net

** University of Houston - Clear Lake
   Box 113, 2700 Bay Area Blvd.
   Houston, TX 77058
   (713) 283-3875
   Fax: (713) 283-3810

*  MountainNet, Inc.
   2816 Cranberry Square
   Morgantown, WV 26505
   (304) 594-9075
   Fax: (304) 594-9088

Abstract

As the World Wide Web (WWW) becomes very popular among internet users, an increasing number of public repositories are using the WWW to promote their services. Currently, a sampling of the major repositories ( ASSET , PAL, CARDS, COSMIC, DSRS, and STARS) reveals that they are utilizing flat files written in HyperText Markup Language (HTML) solely. Electronic Library Services and Applications (ELSA) has gone a step further by using the Multimedia Oriented Repository Environment (MORE [1] ). In this paper, we will discuss in detail, specific activities performed during development, transition, and operation of a public repository using MORE.

ELSA replaced AdaNet on August 31, 1994 when the first public access to our new service was granted. The library is the operational part of the Repository Based Software Engineering (RBSE) program which is a NASA sponsored initiative in software reuse. In a timeframe of approximately two weeks, ELSA transitioned its library holdings and accompanying metadata from a monolithic X-Windows based system to MORE. The improved interface employs client/server technology and is accessible through the WWW. MORE is a public domain, metadata based repository tool employing the WWW as its sole user interface. It consists of a set of application programs which operate together with a stock httpd server to provide access to a database of metadata. The entire interface, client browsing and searching, repository definition, data entry and other administrative functions, are provided through stock Web clients.

Repository assets are classified using a collection (topic) and class (type) paradigm. According to their subject matter, they are included in the collections or subordinate collections that best represent domain coverage. The assets are also classified by media or information type through the class approach. Thus, users can view the information from a top-down perspective through the hierarchy of collections or across collections by the hierarchy of classes.

MORE was designed to support this collection and class model. Navigation is achieved through the activation of high-level hypertext links which ultimately lead to metadata or assets themselves. Searching (Natural Language or Pattern Match) is performed against information provided in the metadata. This combination provides users with a reliable and efficient means of accessing a high volume of assets.

Administrative functions are specifically designed to meet librarians' needs. For instance, assets are stored in "developmental" mode which provides a cleanroom environment for the performance of population and/or certification activities. Developmental assets are only available for viewing by librarians. Following the completion of these processes, each asset is promoted to "production" mode and is therefore accessible to the general user population.

Each collection can have one or more groups associated with it which are authorized to access the assets and subcollections making up the collection. Groups in turn are made up of sets of users and other groups; all defined through the librarian interface. Users not transitively a member of a designated group for a given collection will never see the collection, or its contents, through any of the browser or search mechanisms. This mechanism supports the definition of multiple virtualrepositories in a single physical repository, reducing administrative overhead and allowing for direct sharing of assets.

The MORE interface impacts operational tasks by either enhancing or expanding services. Among the innovations afforded are:

• ELSA clients are automatically introduced to new repository features and assets through an electronic newsletter, produced in HTML format and published electronically. This automated distribution and production process guarantees a more cost-effective and timely approach to reaching clients with relevant system information.

• HTML FORM tags are used to provide clients with an automatic feedback mechanism. Since client input is activiely solicited, the addition of a formatted message encourages information sharing. This forms-based interface also grants Web travelers the opportunity to submit an account registration form electronically, thereby streamlining account activation procedures.

• A separate instance of MORE, called the Guest Collection, is available for general perusal. This sample collection and class set includes assets that represent repository holdings but does not require an authorized username/password; access is granted by typing guest/guest at the authentication prompts. The Guest Collection is intended to provide a window to the ELSA repository, its structure, asset certification framework, and general content, and to encourage registration. Both the Guest Collection and the Registration Form are available from a directory page called the Lobby .

• Sources residing outside the repository are invoked by following links, known as Uniform Resource Locators (URLs), which are embedded within the metadata. Assets that are frequently updated no longer present maintenance-intensive challenges since librarians need only point to the file updated at its source. URLs typically point to files in a directory, stored in a variety of image and text formats, or to directories themselves. URLs provide links to any Internet machine and therefore, broaden the pool of potential asset candidates without affecting system size and/or performance.

• The availability of extended resources afforded by the Web and the ease with which these resources can be included and referenced through MORE, support research into the area of acquisition automation. Several resource discovery and information harvesting tools are currently available or in test mode as prototypes. MORE allows librarians to exploit this technology while providing an extensible classification framework.

• Repository status and usage reports are generated by combining and analyzing data from the server access file and from the database itself. These reports predict population trends and gauge the effectiveness of operational procedures. Reporting capabilities are available to librarians only. Thus, Librarians use reports to drive acquisition methods by depicting user demands and behavior.

• Porting to various database platforms is being investigated in order to increase the usage and potential of the tool. This will ensure that an optimum number of system dependencies are overcome.

• Repositories can exchange lists of metadata on those assets they are willing to share, while the actual assets are still stored in the original libraries. Users can browse or search metadata in one library, however, the resulting assets might point them to another library. Reports on the number of metadata accesses can be provided to participating libraries.

• Queries to remote repositories can be stored as URLs of metadata within MORE. Accessing the hyperlink results in the evaluation of the query on the remote server without requiring the users to specify the query.

• More direct interaction between libraries will be possible with version 2 of MORE, which will support distribution of collections across (possibly heterogeneous) servers.

[1] Eichmann, D., T. McGregor and D. Danley, "Integrating Structured Databases Into the Web: The MORE System," First International Conference on the World Wide Web, Geneva, Switzerland, May 25-27, 1994, pages 369-378.