I'm surprised no one has brought up how
difficult it often is to incorporate design changes into an existing project unless it has been developed with such accommodations from day 1. Even a seemingly simple software project can be monstrously complex under the hood, and something as innocent as "this button should be larger than the rest" can mean weeks and months of proofing, coding, and testing. (e.g. Firefox 3's back button)
Whether a change comes in the form of code fixes, documentation improvement, refactoring, design changes, or infrastructural modifications, the burden of understanding the potential scope of damage and doing the actual work is always, always on the contributor. Convincing another to take on the work on their behalf is always a possibility, but it should not be hard to see why design changes proposed through such a channel will often be pushed to the back of the queue, unless for some dire need.
Designer or coder, there is no excuse for not doing due diligence in making sure a contribution is a good contribution.
Looking at the discussion so far, there appears to be at least some consensus that non-trivial effort above and beyond the (hopefully) expected hand-holding must be dedicated to designers in order for them to become good contributors. And, in corollary, the natural conclusion is that it's unrealistic to expect designers to put in the necessary effort to become good contributors on their own.
Following from the above, the more crucial questions, I think, are "How do we convince projects that they should go out of their way to attract designers", and, "Do you really need designers? Really, really, need them?"
