All about Data Dictionary issues
Over emphasis on Description Logics (RDFS, OWL, Inference & Reasoning etc) matters without any actual real-world instance data (e.g., lot's of reasoning over RDF in zip files or local drives).
All about Linking Openly accessible RDF Data Sets
Over emphasis on Instance Data without Data Dictionary appreciation and utilization (e.g., Linked Data instance level linkage via "owl:sameAs").
All about Applications & Frameworks
Here we are dealing with numerous applications and frameworks that inextricably bind Instance Data Management and Data Dictionaries. Basically, an all or nothing proposition, if you want to delve into the RDF Linked Data solutions realm.
Often overlooked, is the fact that the Linked Data Web - as an aspect of the Semantic Web innovation continuum - is fundamentally about designing and constructing an "Open World" compatible DBMS for the Internet. Thus, erstwhile "Closed World" DBMS components such as Data Dictionaries (handlers of Data Definition, Referential Integrity etc.) and actual Instance Data, are now distributed and loosely coupled. Thus, your data could be in one Data Space while the data dictionary resides in another. In actual fact, you could have several loosely bound data dictionaries that serve the specific Inference and Reasoning needs of a variety of applications, services, or agents.
