|
Asset Locator is a low cost, low maintenance scalable solution for managing development resources in the enterprise, which support search and reuse, collaboration, knowledge sharing, and impact analysis. Throughout the software life-cycle, developers produce an enormous amount of software artifacts, or 'development resources', such as design, code, documentation, test suites, and so forth. Asset Locator reaches configuration management systems, file systems, and other locations in which development resources are stored to find these resources, analyze their content, and create a library of reusable components. Asset Locator provides a set of integrated tools that enable developers to search for reusable code, view the inventory of the enterprise servers, and perform impact analysis activities.
How does it work?
Asset Locator uses a set of autonomous crawlers (e.g., WebDAV, ClearCase, File System) to crawl into the enterprise servers/repositories and find (incrementally) new development resources. Next, a set of domain-specific analyzers (e.g., Java, XML, JSP) analyze the discovered resources by identifying and extracting semantic and textual features, and categorizing them into predefined category taxonomies. The analyzed information is then stored in a DB2 database and the textual information is indexed again to enable quick retrieval of resources by keywords at runtime. Once the enterprise repositories have been mined and analyzed, additional analysis takes place to find the relationships and dependencies between resources. This analysis provides a global view of the domain or the "big picture" of how resources fit together.
At runtime, Asset Locator provides a Semantic Search Service over the Web that enables users (or applications) to invoke queries against the analyzed information. Using the Search Service, users can query for any of the features that were extracted by the analyzers, or perform impact analysis activities. In addition, Asset Locator provides both graphical and non-graphical tools for navigating the relationships/dependencies between resources, and navigating the inventory of the discovered domain using a Yahoo-like category taxonomy.
Status
Asset Locator, developed by the IBM Haifa Research Labs, was introduced in VisualAge TeamConnection v303, GA September 99. Since April 2000, Asset Locator has provided the core engine for IBM's developerWorks' Code Finder which enables users to search IBM Redbooks for sample code.
Asset Locator is still under development, with focus on performing impact analysis across enterprise resources. Asset Locator is planned for shipping in 2002 with additional IBM product offerings.
| |
|
