This prejudice is part of a larger one (which is partly a generalization) that it is rare when commentators -- whether philosophers of science or other -- come close to having a good view. Of course, so many in fields are also quite blind to what is going on. Still, most of the good commentaries I know on science are by scientists, and most of the good commentaries on computing are by computerists. And music ... etc.
My perspective on Castells is biased because it was essentially the backbone of my academic training. I had teachers telling us explicitly that Castells' notion of information society was the framework (in the Kuhnian sense) on which we'd build upon. Having that in mind, I see Castells as a sociologist and his work (sociological) as a view on a society defined by information flows (not exactly computing, although computing and telecommunications are key enablers). Still, he keeps technological determinism at bay by considering the impact of geography, territory, matter - the physical (tangible?) dimension.
What I like quite much about his work is exactly this aspect: he presents a vision of a society dominated by the intangible, while still dedicating rather extensive chapters on the geographical asymmetries of the world. This avoidance of technological determinism, the "information is key but place and physicality still matter a lot, as evidenced by real-world data" notion is why I see Castells as a great reference.