60 likes | 145 Views
Responses from Chris Gill WashU. Topic 1: Emerging paradigms for improvements of software reliability. What are the novel technologies that improve software reliability? QoS in middleware (challenging in Ad Hoc Networks) Communication models tolerant of transient connectivity
E N D
Topic 1: Emerging paradigms for improvements of software reliability • What are the novel technologies that improve software reliability? • QoS in middleware (challenging in Ad Hoc Networks) • Communication models tolerant of transient connectivity • How to use (model-based) analysis? • Integrate models and planning • E.g., designing motion and availability profiles w/in mission plans • How to build contingency management into software? • On-line adaptive QoS monitoring and reconfiguration • How do paradigms scale up? • Composition based approach to functional & QoS aspects
Topic 2: Balance between formal and informal methods, engineering and artistry, evolution and rebuild • Is the objective of formal methods proof or understanding? • Yes, both! Defeasible logics, argumentation are needed too • Argumentation may be more powerful than proof in practice • E.g., different views of battlespace from different agents • Argument as a way of finding consensus/compromise • How should models relate to programs? • are models the source code or abstractions of the source? Yes! • do they represent individual traces or families of behaviors? Yes! • Are non-procedural languages viable? Will they help? • Middleware policies/mechanisms as well as languages • CIAO, nORB, TAO • Do visual notations help? Which ones help most? • Help whom? What is the right user interface? • Are source code annotations acceptable? useful? • Definitely – allow low-level source code review • But, do we want to invest in readability for generated code, etc.?
Topic 3. The role of open standards, open- source development, end-user programming and other radically different models • What are the good candidates for open standards? • Simpler ones are better • Crossing the chasm between one and many design dimensions • What is the role of open standards in software reliability? • are there standards for reliability? • do standards improve reliability? • Is there relationship between open-source and software reliability? • Is openness about intellectual property or transparency? • Better programmers write better code • Good and bad examples in open and closed source • What program development models increase reliability? • Hiring the best programmers possible
Topic 4: Effects of software certification on the current balance between software/system modeling, analysis and testing • Given that certification processes require the same degree of confidence in all evidence used therein • E.g., when less direct (e.g. non-system-testing) methods are to be used for certification: • Any models, modeling tools, analysis tools, etc, must be certified to the same level as the underlying system, and • Fidelity of any system representations must be similarly certified for all applicable aspects (e.g. functionality, timing, concurrency), and • Measurement fidelity must be better than quanta being measured • …What unique approaches, if any, can more efficiently certify tools and representations than the system itself? • …What techniques are available which compellingly reduce the aggregate certification effort? • …What aspects are most amenable to these approaches and techniques? • Can techniques address unique challenges result from mobile ad hoc networking or other system of system dynamics? • Yes: reliable message delivery in the face of transient connectivity by exploiting motion and availability profiles. • …Are these approaches and techniques scalable to the size of FCS? Yes, if we can find ways to decompose the problem.
Topic 5: Reliability metrics for real-time, embedded software for FCS • Are reliability metrics for real-time embedded software different than for other kinds of FCS software? • Yes. Time once spent is gone, which differs from other reversible resources like battery charge. Temporal, power, and storage limitations add important new failure modes. • How should the contribution of software to overall system reliability be defined and measured? Is it reasonable to say Rsys = RHWRSW? • Depends on the definition of algebraic operators X, = • What process should be used to confirm that FCS software satisfies high requirements for reliability and availability? • Research and development done on representative testbeds • Requires significant investment up front • how should field experience be used to establish a software reliability estimate? • Similar uses to attack models in security: reveal failure modes