There is inherent bias in academia and unwillingness to accept ideas that go against the canon. That being said the worst case is still scary - crackpots could be given an apparently legitimate platform. The best case is exciting, an alternative outlet for radical ideas that lost the battle to get published. Some reasons I've heard for not being published: the paper is obscure and thus the audience is limited, the paper is obscure and thus it doesn't fit this journal, the author is unknown or it doesn't follow our guidelines. All these types of papers would be perfect for a journal like Rejecta Mathematica.
If you actually look up some of the authors in the first issue http://math.rejecta.org/vol1-num1 , you'll see that some of them are pretty legit scientists (eg Stephen Boyd who has a very nice book on convex optimization on his website)
There is inherent bias in academia and unwillingness to accept ideas that go against the canon. That being said the worst case is still scary - crackpots could be given an apparently legitimate platform. The best case is exciting, an alternative outlet for radical ideas that lost the battle to get published. Some reasons I've heard for not being published: the paper is obscure and thus the audience is limited, the paper is obscure and thus it doesn't fit this journal, the author is unknown or it doesn't follow our guidelines. All these types of papers would be perfect for a journal like Rejecta Mathematica.
EDIT: added last few sentences