- From: Simon Cox <simon.cox@ned.dem.csiro.au>
- Date: Thu, 24 Feb 2000 09:54:55 +0800
- To: www-rdf-interest@w3.org
- Message-ID: <38B48F6F.D4EB6421@ned.dem.csiro.au>
This approach is certainly not a problem for dc, since the DC definition explicitly states that in any description-instance all properties are optional. However, many schema's have more restricted obligations. For example, in the ISO/TC 211 metadata standard (now defined in UML) they provide explicit instructions about how to create a custom "profile" of their base standard. The derived schema (i) inherits all mandatory components, plus non-mandatory ones as required, (ii) can increase the obligation, or restrict domains, and (iii) a single attachment point is provided for extensions. My interpretation of this is that they are requiring you to use the ISO schema as the "base" schema or not-at-all, and more or less prohibit you from extracting anything less than the complete thing for re-use. Pete Cliff wrote: > > I am developing an RDF schema for Collection Descriptions (call it > rslpcld) that is made up of a number of properties from a number of other > schema - for example it is defined in terms of: > > dc:title, dc:identifier, dcq:location, cld:strength, cld:accessControl, > vcard:voice, etc. > > but does not use ALL the attributes in any one schema. > > How do I define the namespaces in the rslpcld schema? My initial thought > is to include a "xmlns:" line for each of the schemas used. However if > this simply "#includes" the properties of other schema into the rslpcld > schema then the rslpcld schema would be defined in terms of a number of > undesirable properties. -- Best Simon
Received on Wednesday, 23 February 2000 20:58:51 UTC