It's probably the former. It's been tried before, with Deep Space Two at mars - but didn't work.

There's also a sense that if you're going to spend tons of money to send a once-in-a-generation probe to an outer planet, it has to work.