Summary of GSAW 98 Plenary Sessions What problems are we solving Reduced budgets desire to reduce time to market rapidly moving technology n Stovepipes don t work in the long term n Cost of software changes is disproportional to changes in requirements n software is not soft architectures don t support evolution What problems are we solving con t Reuse isn t being done effectively n Difficulty in adopting commercial practices n Sequentially engineering space and ground as well as hardware and software leads to insert your worst fear n Lessons learned n One size does not fit all Need systems that are tailored to requirements Share core assets but customize to fit needs n Architectural issues are key to interoperability standards are not enough commonality of user interfaces flexibility evolvability transparency lifecycle costs product product line Lessons learned con t Government needs to open up the architecture to allow industry to provide commercial products n Good architecting requires n Good engineering Expertise in multiple disciplines n Need to identify the degree of variability the architecture needs to support to help in evolving architectures Lessons Learned con t n Product line approaches offer the potential for large scale productivity gains industry has succeeded management commitment build on well engineered well understood core assets need a common vocabulary to help identify common assets Lessons Learned con t Need to build skeletal architectures not throw away incorporating COTS n Autonomy requires an integrated approach to space and ground n Technology can help Architectural flexibility evolution incrementality n Architecture evaluation architecture tradeoffs n Automated generation n Assurance through model analysis n Technology to support automation and autonomy n Organizational issues are challenging We build what we are organized to build n A growing and diverse space community DOD civilian space US commercial space and international space n How to effectively carry out our approaches to reducing cost How can we tell how well we are doing n CCT as a product line DII COE as a product line infrastructure Common HMI Memorable phrases terms Better faster cheaper RISKIER n Tail vs tooth n Stovepiped hodgepodge n Autoshave n We build what we are organized to build n Conway s law If you have four groups working on a compiler you ll get a 4 pass compiler
View Full Document
Unlocking...