Presentations for this Workshop can be found here or on Agenda Link.

Location: Gerontology Auditorium
                  University of Southern California
                  Los Angeles, CA

     The Annual Research Review will be held Tuesday, February 6, 2001. It will summarize our recent results on our MBASE development methodology and experience, COCOMO™ II extensions, commercialization of Easy Win Win on GroupSystems.com's collaboration infrastructure, and architecture research, plus presentations covering our new National Center for Empirically Based Software Engineering (CeBASE) collaborations. A preliminary agenda for the Review is also attached; it also includes our evening Affiliates reception, dinner, and feedback session, where we'll also hand out free copies of our new COCOMO™ II book and CD.

     The CBS workshop will continue our successful series of collaborative workshops with the SEI, and inaugurate a set of workshops based on our collaboration with Victor Basili and the University of Maryland in our new NSF-sponsored Center for Empirically Based Software Engineering (CeBASE). A particular focus will be on how collaborative empirical research (quantitative and qualitative experimental and observational analyses of CBS phenomena) can best be pursued to provide more useful guidelines for CBS development and evolution. Some particular issues that we propose to cover via breakout groups are: - CBS process and methods issues. How do "requirements" best fit the CBS process? How should object-oriented methods be extended to adequately cover CBS? How do CBS glue code inspection techniques differ from inspecting new design and code?

  • CBS architecture issues. How can you best detect and deal with architectural mismatch among COTS components? How can you best analyze properties of CBS at the architectural level?
  • CBS post-deployment issues. How frequently should you "refresh" an evolving system to accommodate new COTS releases? What kinds of COTS "technology watch" strategies are most effective for anticipating and dealing with COTS change?
  • CBS metrics and predictive models. How can you estimate the amount of glue code needed to integrate COTS components? How can we extend and improve our COCOTS estimation model to cover the post-deployment phase?
  • CBS dependability issues. Dependability includes any desired combination of reliability, survivability, security, safety, accuracy, and real time performance. How can you assess the prospective dependability of (combinations of) COTS products? Are there ways to compensate for shortfalls in COTS dependability? How can you preserve dependability across COTS upgrades?

     I hope you will be able to join us at the Workshop. If you are unavailable at that time, I'd appreciate it if you could pass on this invitation to colleagues who would be good candidates to participate. For any further information, please contact our Administrator, Ladonna Pierce (Phone: (213) 740-5703; Fax: (213) 740-4927; Email: ladonna@sunset.usc.edu)

Best regards,

Barry Boehm