I read the new draft of the DCMI Abstract Model – this is a “DCMI proposed recommendation” and the comment period expires in the next couple of days. I wrote a review of the document, in which I said that DCMI should specify an abstract syntax for DCMI metadata, and a mapping from that abstract syntax to RDF graphs, rather than develop an “abstract model”.
It occurred to me that the purpose of DCMI should be to deal with three fundamental concerns of metadata consumers. Firstly, processing metadata in different concrete syntaxes (syntactic interoperability). Secondly, merging metadata from different sources (low-level semantic interoperability). Thirdly, making valid inferences (higher-level semantic interoperability). I argued that DCMI should specify an abstract syntax (this would deal with the syntactic interoperability) and should specify a mapping from the abstract syntax to RDF graphs (this would deal with semantic interoperability, because both merging and valid inferences are already provided by the RDF semantics).
To put my foot where my mouth is, I wrote an abstract syntax for DCMI metadata and a mapping to RDF graphs.