Are we really trying to, for random example, create a protocol for individually and consistently addressing nodes securely within an ad-hoc mesh network of inconsistent size which needs to route requests to one or more gateways to the larger internet, OR are we trying to help average people create their own neighborhood networks where they can share services?
I submit that we're doing the latter, and the hard problem in order for programmers to facilitate it is the former. As a programmer, I stumble across magnificent libraries bitrotting on various parts of the internet that sometimes do things that I wouldn't even have imagined possible, yet no one but hobbyist techs will ever use them until they are integrated onto the backend of something that designers somewhere created to scratch a person's itch, not a programmer's. The big problem: when will a designer ever be meandering around the depths of CPAN or an aside on some obsessed functional programmer's blog?
To be even more extreme, I'd like to see more projects primarily headed by designers or end user types, ones with little to no technical knowledge and a healthy respect for people who do. As a programmer, I pride myself on telling people the best way that things can be done with a computer, not what people should want a computer to do and how they would like the computer to do it.
Seriously, the less I have to think about ui, especially in the context of projecting myself into the mind of somebody who isn't clear on the difference between a directory and a file, the better. Background shit that automagically runs with no interaction? Love it. Something that helps grammar school kids collaboratively research their school projects in a transparently sourced way? Somebody please draw me a picture, tell me where to put the buttons, and tell me what they should do. And even though I've written an awesome algorithm to realtime index and classify the contributions of the individual students and make suggestions about who should be assigned to clarify a portion of the outline based on keywords and library proximity - tell me where to put it, or whether I've been wasting my time and that no one needs it. I don't have to have a big red button on the toolbar. I'm smart, I can certainly use the algorithm for something else.
Sorry for ranting:)
At most companies, a "designer" is a person who by definition doesn't make anything directly useful. They produce documents describing how they think other people should make things. Much of the time their designs only get executed because somebody with power over paychecks tells developers to obey. But most open source doesn't work like that. Nobody can order a volunteer to do things differently.
What open source needs is for developers to understand designing for people. For user-facing apps, the point isn't to write code, it's to make something that people use and love.
To aid in that, we could certainly use help from user researchers, so we get feedback on what's working and what isn't, and get enough data help us think about solutions. And maybe we can use help from consulting designers when we need help with a problem. But I don't think we need designers as such, any more than we need managers.