RDF is "just" a style (and i hope i'm wrong)
recent RDF gatherings (#SemTech and #RDFn) have generated some interesting twitter traffic (twaffic). and reading the Simple (unofficial) results from RDF Next Steps Workshop posted by @sandhawke gives me reason to contemplate (again) the current nature of RDF.
RDF is not...
RDF is a standard model for data interchange on the Web.
RDF does, however, use these technologies. RDF messages can be expressed using a number of registered media-types (RDF/XML, n3, turtle, RDFa, SPARQL Query, etc.); it can be archived and locally accessed using various storage models such as RDBMS, native file systems, etc; and it is most often exchanged over the Internet using the HTTP application protocol.
RDF is based on the idea of identifying things using Web identifiers (called Uniform Resource Identifiers, or URIs), and describing resources in terms of simple properties and property values.
RDF is a set of rules for [en]coding information into messages. RDF itself is not a "standard" but there are a number of W3C Documents related to RDF. RDF has a simple and straight-forward syntax and, as an information [en]coding style, it has much in common with programming styles like OOP, functional programming, etc. just as a programming style can be applied to a wide array of programming languages, the RDF style can be applied to a wide range of MIME media types. for example, in addition to the RDF-centric media-types mentioned above, several versions of RDF-style encodings for the JSON media-type have recently appeared (Talis RDF JSON, RDFj, JSON-LD, JRON).
RDF will possibly never be...
as i scan the topics discussed at RDF Next Steps i notice that the bulk of interest is in improving the quality of information within a message/graph (along w/ the details of merging graphs). i see very little about the details of sending messages efficiently over the Internet and/or enriching RDF media-types with application control data (in order to support Hypermedia Types). this, despite the fact that my impromptu Hypermedia RDF lightning talk at the Linked Open Data (LOD) - W3C Track @ WWW2010 seemed to be well-received.
for the near-term (likely through "RDF 2.0"), it looks as if the issues of optimizing the RDF style for application-level protocols and network architectural styles will have to wait; possibly forever.
you know, that's ok.
there are lots of very successful information interchange formats that lack application control information (JSON, XML, CSV, etc.). it might be time to accept that RDF will never be a control-data rich format like HTML, VoiceXML, SMIL, etc. that RDF, like other non-Hypermedia Types, will always be dependent upon custom-written stand-alone applications, code-on-demand, and plug-ins in order to interpret the raw data. that the RDF style will never have a powerful, user-friendly viewer like the common browser; one that understands Hypermedia types; that allows content providers to enrich the data in ways that guide humans and machines to the level of dynamically executing applications just by loading and parsing marked-up content accessible via a URI. a single format that is both data and application code.