RE: Mapping between schemas

I don't think there is a uniform approach because the original specs aren't
uniform in use of a "qualifier".

Looking at vra3:title:

vra3:Title.Variant
    could be subProperty of title
    its still a title for the work
vra3:Title.Translation
    could be subProperty of title
    its still a title for the work

but

vra3:Title.Series
   Not a subproperty
   Would seem preferrable to link to the "series" description
vra3:Title.LargerEntity
   Not a subproperty - this isn't a title for the work

In Dublin Core there is explicit subProperties:

dc:created     subPropertyOf  dc:date
dc:references  subPropertyOf  dc:relation
dc:medium      subPropertyOf  dc:format

and if the subProperty is a true statement so is the same pair related by
the superProperty.

The var3 mappings to DC also need thinking about 

vra3:measurements is defined to map to dc:format
  measurements.{dimensions,format,resolution}
is about the image (actually about the work or about the image)

vra3:material is defined to dc:format but is about the substance of the
work.

so there seems to be confusion between the 

	Andy


-----Original Message-----
From: Butler, Mark [mailto:Mark_Butler@hplb.hpl.hp.com] 
Sent: 7 October 2003 17:40
To: www-rdf-dspace@w3.org
Subject: Mapping between schemas



Hi team,

After writing an XSLT transform to turn the Artstor XML into RDF/XML, I
decided to have a go at writing an RDFS schema for the resulting RDF. I then
decided to try to link this schema to an earlier one I had done for VRA
Core, because the Artstor metadata is based on VRA Core.

There are a number of problems doing this, but I came across one which I
want to mention here because I suspect it may be potentially very generic.
It is the problem where one schema uses properties whereas another schema
uses classes.

For example, consider two schemas that both use the VRA element creator that
refers to an image record. Note I am not using RDFS class / property
terminology here deliberately, because specifications like VRA Core do not
use such terms. Creator has a number of qualifiers, e.g. Creator.Role,
Creator.Attribution, Creator.Personal_name, Creator.Corporate_name. So how
do we represent this? Well there are two approaches:

(a) We can create a property called creator, and then subproperties called
role, attribution, personal_name and corporate_name. If we do this, we are
making the assumption that an image has exactly one creator.

(b) Alternatively we can create a class called Creator. Now our Image
instance has one or more properties called hasCreator, each of which points
to an instance of Creator. The properties roles, attribution, personal_name
and corporate_name all have domain Creator. Now images can have multiple
creators, because each creator is an independent object, rather than a
property value. 

Now lets consider the mapping:

1. It's fairly straightforward to map a:role, a:attribution, a:personal_name
and a:corporate_name onto their respective counterparts in b.

2. Mapping b on to a is may be more difficult, if an image does have
multiple creators.

3. Mapping between creator is difficult, because it is a property in a and a
class in b.

any thoughts?

Dr Mark H. Butler
Research Scientist                HP Labs Bristol
mark-h_butler@hp.com
Internet: http://www-uk.hpl.hp.com/people/marbut/

Received on Wednesday, 8 October 2003 06:50:37 UTC