I am in Paris

Launch Service Cup 2010

It is my third year to chair Service Cup. In 2010, the theme is “Applications and Algorithms on Clouds”. We want to see business applications as well as algorithms run on commercial or private clouds. As clouds software and hardware are available now in many ways, it is possible for universities and companies to launch clouds project. I should say, it is also a learning process for me to this new topic. However, I believe it is an interesting and promising direction for business as well as research. The difficulty for us is to attract participants. I am afraid that not many universities or professors are working on this yet. However, to be the first to do it is interesting.

Reborn of Computer Labs

A chronicle article proves my idea – the computer labs used for teaching need to reconfigure themselves in order to be better used. Most students, over 85% as the article says, have their own laptops. The pattern of the usage of the lab is changing. Students can work on their own laptop at a more comfortable places like in their own bedroom or even in their own bed. The lab becomes a gathering place for discussion. Students meet in the lab to discuss and synthesize their project work, present their work to the others. For the sake of training, I also encourage the students to do a full development cycle from installing the development environment to testing and deployment. Sometimes, this can only be done on their own machine. For example, our lab discourages the students to install certain type of software. Also, in a lab, the development environment is set up for them. They can jump into coding directly. It is enjoyable, however, they will find they are left alone if they do not know how to set up the environment at workplace.

Second, the lab should provide something beyond the capacity of their own laptop, for example, expensive software and powerful computation resources. They should be able to use them from remote access from their own laptop. I find the students do not fully use what we have invested in computation resources. Even for research like machine learning and data mining, the graduates still do their experiment on their own laptop or PC. An experiment can take several days. They should be trained to write concurrent algorithms for their algorithm. Or simply submitting their jobs to a powerful multicore computer can save their time.