A concise synopsis of the chaotic situation that is present in information technology is this:
The user is experiencing a threefold Tower of Babel in semantics, i.e. conceptual modelling, in computer programming, i.e. symbolic modelling and in persistent or temporary storage and retrieval of information, i.e. data modelling. Is it possible to bring order to that chaos ?
My answer is yes provided that we are tuned into a similar perspective as that of R3DM and Wolfram Technology. The reason I am referring to Wolfram technology in that post of mine is to observe and acknowledge that the combination of Wolfram Alpha, Wolfram Language and Wolfram Kernel demonstrate a primary implementation on the three levels of abstraction that I have defined in R3DM (Resource, Representation, Realization).
I have been inspired to formulate R3DM when I started reading publications on the "identity crisis" problem that is still evident in the "Semantic", i.e. linked data, RDFised, web four years ago. The interoperability of computer systems in healthcare was another reason and a third one was the personal information management of information resources. Thankfully the maturity of Wolfram technology, and to be precise the immense research of this product in many fields of computer science for more than three decades of research, came just in time to provide the means to enter a new era of information technology.
R3DM can be a concise perspective on the operations that take place in Wolfram Techology. To begin, RESOURCES at S3 (Signified) level, may ORIGINATE from Wolfram Alpha. They can equally originate of course from Wikidata, Freebase, OpenCyc, or any other semantic database. This is all about conceptual modelling, i.e. concepts and their associations. Then it comes the symbolic manipulation, in my opinion this is where Wolfram technology excels. It is a unique blending and style of functional programming together with rule-based programming.
We have already entered the symbol level S2 of R3DM. Therefore information RESOURCES can be REPRESENTED THROUGH Wolfram Language in ANY WAY you can imagine, think about numeric symbols, graphics, sounds, images, etc. Everything is there ready for use. I think what most people find hard to conceive, is the connection of the two levels S3 and S2. After all both "Concepts" and "Symbols" are equally represented with symbols. The hard part is that any concept can be represented in numerous ways, with any kind of symbolic expression speaking in Wolfram language terminology. Nevertheless the semantics, i.e. meaning of concept remain the same, whatever representation, i.e. syntax, symbols you are going to use. In my opinion there is a lot of work to be done in the connection between these two levels and the modelling of the highest level S3.
But to continue the story, symbolic expressions at level S2 have to be encoded, i.e. see REALIZED in R3DM, at a lower level S1 (signifier). This is the land of data encoded in any form including binary forms. Think about JSON, XML, HTML, RDF, PDF, JPEG, XLSX, BIN, TXT, etc. This is another major stumbling block because any symbolic expression can be encoded in myriads of file - data serialization formats. Even worse, in my opinion very often developers cannot grasp the fundamental distinction between symbolic expressions and permanent/temporary storage i.e. the distinction between symbols and data. In R3MD data are mere content and containers with structure, i.e. data structures. Wolfram Language has a lot to exhibit here in terms of data transformation and importing, exporting but the level of data has not become separable.
You are right to think that the border lines among the three levels S1, S2, S3 I described are bleared. But I think the challenge of any system or framework is exactly that to define precisely the boundaries and make clear and distinct the operations at each level. This is what I am working on.
Read more and comment at Computational Semiotics group, http://goo.gl/8ZuikV