From W3C Wiki
Jump to: navigation, search

A representation of one possible architecture for the Semantic Web The semantic web architecture is not universally agreed. The exact order, names, and status of the layers is under discussion around the community of researchers, developers and other interested parties.

This article may turn into an examination of how this architecture is useful, effective or valuable. It is something of an experiment to see if WikiConsensus around the (arangement of the? definition of the?) layers emerges.

Some components of the Semantic web architecture that already exist are (in no particular order?)

  • RDF, which captures the "existential conjunctive" fragment of logic
  • conjunctions of these statements
  • DereferenceURI mapping from URIs to RDF graphs.
  • syntax specifiying permissible structures for representation of data.
     The RdfXmlSyntax  standard is the basic standard,
     other serializations such as NotationThree exist.
  • vocabulary for describing properties and classes (of data items?)
    • specialization and generalization of properties and classes may be achieved with RDFS (RDFs?)
    • Relations (cartesian pairs) between sets (one-to-many, many-to-one not many to many?) and set operators (union, intersection) ... Web Ontology Language (OWL)
  • queries (RDFQueryTestCases) and rules, using universal quantification.
    See SPARQL which is the standard
    query langauge and protocol.
  • digital signatures (and canonicalization of XML)
  • ProofChecking ensuring that 'explanations' (specified as queries?) of assertions can be verfied (using trusted syntax specifications, trusted vocabulary, trusted RDF statements (and other things?) )
  • RDF Forms to bind the hypermedia application model to RDF

Deployment of the Semantic Web's Architecture

The original deployment of the web was fueled by the replacement of existing hypertext systems with web-based ones, a process known as 'webizing'; i.e. replacing all the identifiers by URIs. This allowed references within and across documents and within and across organizations. In turn, this allowed AnarchicScalability, because the web allows indefinite addition of nodes.

Likewise, SemanticWeb Architecture is what you get when you take databases and knowledge representation and use URIs to represent aspects of them. These, hopefully, will relax the classical ClosedWorldAssumptions of those working with, e.g. DBMSs, which will result in similar effects (to 'webization') in the representation of the physical world, its states and behaviours.

Deployment of SemanticWeb related technologies could be accelerated by

  • the maturing of research
  • applicable research
  • ratification of standards

It could be retarded by

  • vendor opportunism
  • useless and irrelevant research mistakenly deemed significant
  • patenting of useful mechanisms as they are discovered

Some sticky bits include RdfAndMediaTypes.

The gap between mature research (mature in what way?) and emerging standards (what use are standards if they are not ratified and adhered to?) is closing (what are you talking about?).

As the time taken to get from mature research to development of standards decreases, the deployment of SematicWeb will accelerate (how can it be deployed if it's still in development?).

The Semantic Web Bus

The SemanticWebBus is the pivot point of interoperability that we're (who? The W3C? Anyone else?) aiming for.

Please define this term here or on the article page-- this is something Tim Berners Lee has talked about, it seems, but the only thing I can find is this slide.

A Semantic Web Programming Language

Java as a programming language lames for the use in Semantic Web. A language like prolog, datalog, or APL would fit more the Semantic Web, a suggestion for a "perfect semantic web programming language" is here, please contribute:

Please see also the /RefList

Unsorted Text and Discussion

hmm... what's the PPR:OpeningStatement of this topic? i.e. the arguable thesis?

  • One aspect seems to be "what would cause the adoption of the SWA in the way the web was adopted?".(I would say SW-related technologies, myself-- there is no SWA, is there? It seems a bit prescriptive) guest

hmm... layer cake/stack/wave diagrams (LinkMe) are fairly memorable. LinkMe?

  • Que? guest

"It uses URIs for naming, XML for interchange syntax, and RDF for basic facts and descriptions."