CUSEC Day 2

Posted on January 16, 2004
Filed Under Software Eng |

Michel Dagenais, a professor from Polytechnique began the day with an informative session discussing Free Software. For some it the crowd this was new information, but it is always good for new people to see the light. Some insightful points made were about the future of Free/Open Source software. There were are: mature infrastructure areas are almost entirely running on FLOSS so there will be a shift to specialised, and quickly changing apps because of the development model’s adaptibility. There will be a move from generic libraries to large architecture technologies. And big inrods in to the coming boom in the wireless domain.

Dr. Lionel Briand, Canada Research Chair in Software Engineering at Carlton University, discussed software dependability related to quality. His focus was Fault Avoidance, Fault Detection, and Fault Tolerance. He did not discredit rigorous methods, only said that no matter how rigorous there will be still faulty software. An true statement until code generation is automated. These faults can only be discovered by testing and inspection, and Dr. Briand spent most of his time talking about testing. He applied the interesting testing methods to real world topics, and also gave some insight from his first hand experience on the team that was responsible for discovering the error that lead to the Ariane 5 disaster. His conclusions where to apply optimization and search techniques to automate testing, I guess I should get back to SFRW 4TB3.

The final talk of the day was a tutorial session with Peter Hitchcock, from Dalhousie. He showed how to formalise specifications in Z. He used the (I guess classic example) that Dr. Khedri used in SFWR 4EF3. Nothing new, but I had a problem that exposed an ambiguity, that is a result of the discrete math rules. But when applied to specifying a software system it can lead to problems in later stages. This problem with Z was not made evident in 4EF3. I asked Dr. Parnas about it, and he said that Z is only less undefined that UML, but is still not well defined.

The day ended with cocktails at Newtown (Jacques Villeneuve’s restaurant/bar). The free drink was nice, but when we had to pay, it was a bit much for a place that looks like every other bar in downtown Toronto. When then when for Thai and had a good meal, but we were there forever. Had fun with the Waterloo SEs we ate with. They are a great bunch and we should have them down to MAC sometime.

Comments

Leave a Reply