CUSEC 2004
Posted by Skrud at Saturday, January 17th 2004 at 10:47pm
I spent the last three days straight at the Centre Mont-Royal attending CUSEC (Canadian Undergraduate Software Engineering Conference). This was the 3rd annual conference and had speakers from across the Software industry presenting as well as students from across Canada attending. The theme of the conference was focused around “quality” and what it means for Software Engineering. How do we create “quality” software? What is quality when referring to software? How do we address the issues of bugs in software? How can we prevent them? So on and so forth… There were some very interesting presentations on these subjects as well as a neat demonstration from Apple on their RAD tools. There was a also a programming competition sponsored by Nullsoft and some members from Team Nullsoft were actually around to judge it. (Yes, these are the same guys that wrote Winamp.) Unfortunately, I wasn’t able to compete in it. :(
Some big names showed up to present, including Fog Creek’s Joel Spolsky who presented the most entertaining of all the lectures at the conference. He demonstrated how one can view software like an iceberg. Some 90% of the iceberg is underwater and can’t be seen from the surface, and these are representative of the guts of your program: the algorithms, data structures, mechanics, etc. Meanwhile the 10% that’s above the surface is little more than “lipstick,” what the user sees. You could have a perfect program that does everything the customer wants it to do, present it to them, and get a response like “it just doesn’t look slick enough.” Joel also spoke about how users need to feel “safe” the same way they feel safer in an SUV… implying that a user-interface change can make crappier software appear to be of better quality because it feels safer to the user (and he gave Windows XP as an example ;)).
There were also some pretty cool presentations from Ecole Polytechnique professors. One, by Pierre Robillard, went through how software could have worked for a herb trading guild in the middle ages and examined how they came up with bugs and how they could hypothetically deal with them. There was also a pretty cool speech on Free Software and achieving software quality in open-source domains.
Many of the discussions however ended up being about software specificiations. In order to write quality software, clearly defined requirements need to be set up and complete specifications need to be written in order to ensure software quality. Interestingly enough there seems to be a lot of disagreement in this area. One professor from the University of Waterloo spoke about how Domain Specific Languages (specification languages engineered for specific purposes, ie, automotive engineering) can help by providing highly specialized detail in a highly specialized way. We were also presented with a specification language called Z by another professor. Meaniwhile, David L. Parnas of the Software Quality Research Laboratory in the University of Limerick, Ireland disagreed completely (and eloquently) with both of these ideas saying that “Z is bad” and that the multitude of Domain Specific Languages is a sickness that software engineers seem to contract. Needless to say, this made for a very interesting panel discussion today.
Anyway, I’m definitely looking forward to attending CUSEC next year (even though it may be in Ottawa by then) and competing in the coding competition this time so as to defeat the guys at McMaster University. (And I’ll do it in KDE too. :P)
