Other things being equal I tend to link to Wikipedia pages for definitions and stock proofs, in part because I've been inspired by researching past questions here to fix things there, and it seems to me as likely such cross-linking increases probability of articles to be maintained or improved.
However I have had a couple of qualms in the past week about such favoritism. Sometimes a Wikipedia page is really only a stub. Even when it defines a required term adequately, I'm not sure how much allegiance is really called for.
I'd like some discussion of this strategy of favoring Wikipedia or some other sites. I wonder if the community ads, say, for PlanetMath should be considered (although they went through a rough patch not long ago). I would think we'd favor sites like Wikipedia and PlanetMath that agree with community sourcing of material, and by the same token I'd be reluctant to give extra weight to linking Wolfram's MathWorld (acknowledging that we have community ads for Mathematica and that sometimes the MathWorld articles win by dent of having the best exposition).
What do you think?