 |
 |
On Wed, 12 Aug 1998, Mr C A Rusbridge wrote: > It seems to me that the metadata alluded to in this message is straying > too far from the goal of Dublin Core as metadata to aid discovery. Surely > this is entering the domain of preservation metadata? > > We must avoid trying to pack everything into DC and remember the Warwick > Framework approach (now being instantiated in RDF?) of having several > packets of metadata for a resource, where each packet is specialised for a > particular purpose. Quite. The big challenge comes when two domains overlap. WF's approach to modularity allowed each vocabulary to be represented in a different syntax. RDF imposes tighter restrictions; all vocabularies are expressed within the same data model (and, typically, syntax). This takes the spirit of WF but suggests possibilities for a finer grained intermingling of vocabularies -- bought in return for forcing a common model/syntax on the world. This means that a particular chunk of metadata can have multiple vocabularies in scope at once; eg. DC specialising in resource discovery, vCARD for information about people and so on. Since these domains overlap, it'll often be the case that there are similar constructs in both vocabularies. This makes the idea of having a vocabulary that gave relations such as THES:broaderTerm, THES:narrowerTerm, THES:synonym etc quite appealing. RDF itself gives us the RDFS:subClassOf relation which lets us say that a category in one vocabulary is a subset of a category in another. Since the world doesn't itself fall into discrete and isolated domains, this seems like a promising evolution of the Warwick Framework philosophy: we combine a modularised approach to managing domain-based vocabularies with a data model that facilitates statements about how vocabulary items may be, for example, synonyms. So DC's simple notion of an agent could be hooked up to vCARD's richer category, or DC's simple notion of geography be mapped into richer systems: DC as an "Interlingua"... Dan > > Simon Pockley wrote: > > We are particularly interested in ideas for describing the syntax for operating > > system requirements. In some cases system information is probably more > > important than individual Type formats. > > > > If our content creators describe the system on which the digital resource was > > created. > > > > <META NAME="DC.Format.System" CONTENT="Windows NT 4.0"> > > > > they are providing valuable information about the encoding source of the data. > > Would this be better expressed as? > > > > <META NAME="DC.Source.System" CONTENT="Windows NT 4.0"> > > > > Because we also want to know what is needed to display the resource. > > > > Example: <META NAME="DC.Format.System" CONTENT="Netscape 2 or better"> > > > > Any suggestions? > > > > > > > > -- > > > > -------------------------------------------------- > > Simon Pockley - Cinemedia (http://www.cinemedia.net) > > 3 Treasury Place Melbourne Australia [voice] 61 3 96511510 > > ([log in to unmask]) > > > > > > > -- > Chris Rusbridge > > Programme Director, Electronic Libraries Programme > The Library, University of Warwick, Coventry CV4 7AL, UK > Phone 01203 524979 Fax 01203 524981 > Email [log in to unmask] > >