|
|
AS OF 7/4/2008 8:25PM EST
|
February 15, 2004 —
Software quality analysis-another of the highfalutin terms to evolve from bug tracking-is finding a new advocacy group in the unlikeliest of places: the Architecture-Driven Modernization task force working under the auspices of Object Management Group Inc.
Originally conceived as a way to transform legacy systems for reuse in more modern code platforms, the modernization effort is designed, first, to give organizations a better understanding of their codebase, according to task force co-chair Djenana Campara, who also is chairperson and CTO of Klocwork Inc., which spun out of Nortel Networks Ltd. in February 2001. Klocwork offers software intelligence solutions for analyzing software and design. From that understanding of the code, organizations can then do multidimensional analyses of what they have, and clean it up and set up rules to make sure any errors don't happen again.
"You can talk of defects from a number of different views," Campara said. "There can be architectural defects, defects in the interfaces, in security, in the code. Then there are the inefficiencies and interdependencies that must be understood. Then you can plan how to attack each of those, and to protect against it. You can spend years cleaning up your code, but without protections, within six months, the defects all are back."
Currently, defect-tracking tools tend to be point solutions, Campara said, although "everyone wants development, management and information driven from the same repository." There is not one vendor who can extract everything that teams within an IT organization want to exchange, she added, but under the Architecture-Driven Modernization effort, a metamodel expressed in UML can be created that each team can use that sets the standard for exchanging information via the XML Metadata Interchange (XMI) protocol.
Campara said she envisions vendors of tools for defect tracking, metrics and security, for example, all populating the same repository and sharing the system's information. "When you add data-mining engines on top, it becomes very powerful" for analyzing software for quality, she added. "With so many tools now generating their own repository of information, it's so painful to pull a bit of information from here, and another bit from there."
The downside, she acknowledged, is that a solution of this nature is "expensive at this time."
|