The third day had two astonishing new revelations for the BOINC world. At first there where some talks about virtualizing the scientific application through different virtualization techniques. After the coffee break Daniel Lambrana presented the first BOINC client pre-installed on a USB-Stick that can be distributed by a project and can directly contribute the unused computer cycles of the computer it gets plugged into. There is a Windows and Linux version installed that work on Windows XP and almost all Gnome driven Linux distributions. Windows Vista is currently not supported because the autostart required for this setup is not working. The prototype is connecting the user to the extremadurathome.org project that also has a good video explaining Volunteer Computing.
After that there was a presentation about a Campus-wide Supercomputer at the Pontifical Catholic University of Peru. Genghis Rios Kruger presented the way he and his group of computer scientists used BOINC to build a Grid using the approx. 500 computers of the different computer labs. Interesting here was that they already provide a simple web-driven interface for the scientist to generate work and receive the results. This is going to be made generic and hopefully can be implemented into BOINC or at least made Open-Source so other projects can use it too.
The second revelation was made by Michael Weber from the University of Marburg who is going to publish a new project focused on RNA research, in cooperation with the Rechenkraft.net non-profit association. This is not going to be a university-funded project but purely operated by volunteers.
During lunch break we where at the same place like the first day and it was great. In case you visit Barcelona and need a good place to eat: try the Carmelitas located in the Carrier del Carme.
In the afternoon there where some BOF sessions (Birds of Feather) discussing some issues regarding the BOINC framework and how it can be made more general and how to implement new stuff like virtual applications and easy to use interfaces for scientists. The main conclusions can be found on the wikipage, where the slides of the talks are also published (some are missing but should be added later).