OPACs are, or were, simply stand-alone computers that allowed those in the library to access to the electronic catalog that consisted of MARC records. I was simply surprised that you did not know what they were, which I see as a deficit to our profession. One must know about the past in order to move forward. Of course, I understand the basic nature of  SPARQL, OWL, and HTTP. 

What I find fascinating is your statement that "Linked Data is about modeling the real world, the abstract concepts and physical things and relationships between them, not modeling MARC structures."  MARC is about representing the information resources in a library collection---into which we would also put representations of intellectual content (concepts? Ideas? subjects?) and how they relate to intellectual content of other resources in the collection, as well as represent physical attributes of the resources. And I agree that it can not impose its structure on the real world but neither can Linked Data. But on this I will do some more reading. It seems we've exchanged one technology for another (the former passing on its collective data to the latter),in order to reached a goal that keeps getting more 'defined'?.