 |
 |
As part of the Harmony project, I've been doing some work on combining the semantic knowledge stored in a general metadata term thesaurus/ontology (known as MetaNet and represented as an RDF Schema) with XSLT to enable mappings between metadata descriptions from different domains: http://archive.dstc.edu.au/RDU/staff/jane-hunter/www10/paper.html This paper may also be of interest to the SCHEMAS project. It proposes that each domain should register both an RDF Schema (defining the vocabulary semantics) and an XML Schema (defining the recommended application-specific encodings) and describes methods by which these two schemas languages can be made to work together, albeit not very smoothly or elegantly. jane > On Thu, 25 Jan 2001, Seth Russell wrote: > > > "Sigfrid Lundberg, Lub NetLab" wrote: > > > > > Euhm. What do you mean by "discourse level of dialogue"? That a dialogue > > > cannot take place unless a given statement possesses both semantics and > > > syntax? > > > > No, that dialogue cannot take place unless statements pass between agents. > > I see and agree. > > > > Any XSLT processor will do > > > > I see XSLT as transformations of syntax only ... i can't find the semantics > > there at all ... but then i haven't really studied XSLT yet either. > > You're right, XSLT is for syntactic transformation. But Dan Conally has > written a RDF parser in it :) producing triples out of serialized syntax. > > We've got three things we might want to do with RDF (well, there is > presumably more things, but anyway): > > 1. GET (i.e., http GET) and parse RDF and store resulting triples > in a database. There is a fair amount of software that can do that > by now. > > 2. SEARCH a triple store, GET resulting triples from it, and > PRESENT them as serialization syntax. There are software tools > that can do that. > > 3. Load metadata semantics (i.e., RDFS) into a metadata creation > tool, and create syntactically correct metadata. _I_ do not know > how to do this, and the only thing I can figure ut which is to use > some kind of schema language in order to formulate the syntax that > supports the desired semantics. Schematron is the least painful > alternative, as I see it. > > Any good idea on how provide a general solution to task 3 is appreciated. > > > Sigge