But is Debian an umbrella organization? :P
A big part of answering the question "should this be a Debian project?"
is practically relevant when it comes times to dispersing resources.
Without getting into a debate about how much we should ask GSoC for, I
will say: When it comes to decision time, do we want to use Debian
resources on something that is only tangentially related to Debian, or
does not demonstrably add to the project? I think it's really hard to
justify doing that, and my intuition is to not.