In this paper we take a view from the bottom to RDF(S) reasoning. We discuss some issues and requirements on reasoning towards effectively building Semantic Web Pipes, aggregating and patching RDF data from various distributed sources. Even if we leave out complex description logics reasoning and restrict ourselves to the RDF world, it turns out that some problems, in particular how to deal with contradicting RDF statements and patching RDF graphs, do not yet find their proper solutions within the current Semantic Web Stack. Besides theoretical solutions which involve full DL reasoning, we believe that more practical and probably more scalable solutions are conceivable one of which we discuss in this paper. Namely, we provide means to express revocations in RDF and resolve such revocations by means of a specialized RDF merge procedure. We have implemented this conflict-resolving merge procedure in the DBin 2.0 system.