I don't disagree, I just don't want to volunteer projects for something they don't want.
Right, there must be clear indication that they are willing to accept the work when it's done.
Note also that some of the largest Python-based projects, Django,
I have a working port of Django to Py3k, however, the Django authors are not interested in integrating it. The same could easily happen to other ports. I can accept that, and just wait a couple of years until they are ready. The GSoC student can't really wait that long.
OTOH, it could be part of the student's application to get in contact with a potential mentor, and we could prioritize porting projects assuming the package authors indicate sufficient intent to accept the results of the porting.