As the research process itself should be as transparent and reproducible as possible and as it is increasingly relying on the computational analysis of large data sets (depending on the research field) open source software should be applied as much as possible in science. Due to this it is very important that future researchers/engineers are trained in those tools and get aware of the concepts of open source software. Luckily there is a trend that this is happening. Institutions move away from teaching programming based on proprietary software like MatLab and instead build on open solutions like R (for statistics) or Python (as a general purpose language) using the IPython notebook.

http://www.r-project.org/

http://cacm.acm.org/blogs/blog-cacm/176450-python-is-now-the-most-popular-introductory-teaching-language-at-top-us-universities/fulltext

http://python.org/

http://ipython.org/notebook.html

I whole heartedly second Dr. Förstner. More open source softwares should be used in scientific research. A regulation could also be put forth in future that no research softwares should be proprietary. In many cases, particularly in the third world countries, only the trial versions are available for researchers, in which the most useful tools are always locked. This is not a healthy trend for quality research in science and technology.

Totally agree, using OSS, no matter the area comes with the opportunity of being enhanced, a team can code, a team can test, a team can improve. This feature of code is  spreading over the work, and results are welcome for obtaining better products and tools.

FOSS has the posibility to be a peer reviewed software. You also can reuse programming code from others and not re-invent the wheel again and again