My main concern is that who has the expertise to answer the last two questions?

Funnily, the answer for Moo! is yes for the both questions. It has contributed to a field of research (there are peer reviewed articles on scientific journals about it) and it has historical and ecosystem building value regarding distributed computing and the huge effort done to break a key this hard. Prize and charity is a nice small bonus too.

The problem with BOINC projects overall is, that topics that are deemed scientific enough by scientists, will usually get funding to get worked on by a supercomputer. Thus, BOINC projects will often tend to be brute forcing or borderline scientific projects.

If Gridcoin really takes off, we will really want there to be enough projects to choose from and have work units available. Currently team gridcoin is outputting 5000 TeraFlops, while bitcoin mining is aproximated to be at 82 zettaFlops = 82 000 000 000 TeraFlops. I think Moo! could fit in there.