I doubt there's much negative feelings here regarding applying the open source way to science, but I'd like to mention the prevalent use of educational discounts, which has a terrible side effect of getting students hooked to specific software which they will then "need" when starting in a commercial job.
In the Netherlands, this way of discounting to the point of giving away to students of whatever level has turned the country in a 99% windows place.
Similarly, the use of matlab in science, which is more or less affordable in the academic world, makes it attractive to build software on top of it in that world, but hard to include other, less fortunately situated people to work on that software (e.g. http://fieldtrip.fcdonders.nl/).
So this is just the first step, they should further require that no proprietary software is necessary to reproduce the results. Such a rule would probably motivate (i.e. get money for) the people from FieldTrip to port it to Python or Octave, whichever is easiest.