If you don't think that citations are important, that puts you at odds with every academic institution in the universe.
I don't know the exact reasoning why primary sources are needed... I suspect that the reasoning is that if you're interpreting an interpretation, eventually you are playing telephone with the original research. Encyclopedias typically do not cite sources, so it is not possible to paper-trail your way back to who it was that actually discovered the information.
There's an additional layer of "terrible idea"-ness for citing wikipedia: the information there is dynamic. I think that's the best case for excluding it as an academic source. If I say that the Old Man Murray forums became Caltrops.com, and cite wikipedia as my source, what can you say about that? It *did* say that at one point, and I know it to be true, but it does not say that now. If you delve 3 years deep into the revision history, you can see places where it said that. That's too much to ask of a fellow researcher, in my opinion - citations need to cite sources that aren't dynamic in that regard.