Hi Rachel,
> The DCAM model defines a description set as:
>
> 'A description set is a set of one or more descriptions about
> one or more resources.'
>
> I think it would be helpful to model 'description sets' in
> more detail (not necessarily in the DCAM document, perhaps as
> an additional document). It is not clear to me how the
> different descriptions in a description set are distinguished
> - I cannot see any indication in the DCAM as to how multiple
> descriptions of multiple resources are differentiated. I
> would expect the relationships between descriptions in a
> description set to be 'modelled' in some way.
They are differentiated by the fact that a description describes exactly
one resource.
The DCAM doc says:
===
A description is made up of one or more statements (about one, and only
one, resource) and zero or one resource URI (a URI reference that
identifies the resource being described).
===
Now, yes, there is work to be done on how to express this in the various
syntax specifications (and the "zero resource URI" case sometimes
worries me...)
> I have missed recent f2f discussion, so perhaps I am misunderstanding!
>
> As I understand the DCAM definition of a description set, it
> suggests descriptions can be grouped into a description set
> in an arbitrary way. Am I right in thinking there is no
> necessity to relate the descriptions together in any structured way?
Consider:
(1) a set of descriptions consisting of two descriptions of two persons
(where those descriptions do not include any statements expressing
relationships between the two persons)
Description
Resource URI : http://example.org/person/TM
Statement
Property URI : http://example.org/term/name
Value String : "Thurston Moore"
Statement
Property URI : http://purl.org/dc/elements/1.1/type
Value URI : http://example.org/terms/Person
Description
Resource URI : http://example.org/person/LR
Statement
Property URI : http://example.org/term/name
Value String : "Lee Ranaldo"
Statement
Property URI : http://purl.org/dc/elements/1.1/type
Value URI : http://example.org/term/Person
(2) a set of descriptions consisting of a description of a group, and
two descriptions of two persons (where the descriptions of the persons
include statements expressing a relationship between the persons and the
group)
Description
Resource URI : http://example.org/group/SY
Statement
Property URI : http://example.org/terms/name
Value String : "Sonic Youth"
Statement
Property URI : http://purl.org/dc/elements/1.1/type
Value URI : http://example.org/term/Group
Description
Resource URI : http://example.org/person/TM
Statement
Property URI : http://example.org/term/name
Value String : "Thurston Moore"
Statement
Property URI : http://example.org/term/isMemberOf
Value URI : http://example.org/group/SY
Statement
Property URI : http://purl.org/dc/elements/1.1/type
Value URI : http://example.org/term/Person
Description
Resource URI : http://example.org/person/LR
Statement
Property URI : http://example.org/term/name
Value String : "Lee Ranaldo"
Statement
Property URI : http://example.org/term/isMemberOf
Value URI : http://example.org/group/SY
Statement
Property URI : http://purl.org/dc/elements/1.1/type
Value URI : http://example.org/term/Person
I think you are asking whether both (1) and (2) are both description
sets.
The answer (IMHO) is that, yes, both (1) and (2) are description sets.
The DCAM doc says
==
real-world metadata applications tend to be based on loosely grouped
sets of descriptions (where the described resources are typically
related in some way), known here as description sets.
==
N.B. "typically". It does not _require_ that the resources described by
the descriptions within a description set are explicitly related. So in
(1) above the two descriptions express no relationship between the two
persons, but the descriptions still constitute a description set.
> In which case, how would one tell the difference between a
> 'description containing statements about more than one
> resource' (which is 'non-conformant' with the DCAM) from a
> 'description set' (which is a grouping of 'descriptions' or
> statements about multiple resources)?
A 'description containing statements about more than one resource' isn't
a description; not even a "non-conformant" description ;-)
> Also I think it would be helpful to define the logical
> relation between a 'description set' and 'related
> descriptions'. (maybe this would be something like: related
> descriptions always form a description set, whereas a
> description set may contain unrelated descriptions?)
A description set _may_ consist of descriptions in which there are no
statements asserting relationships between the resources described e.g.
my case 1 above.
> My feeling is that a description set needs to be structured
> by establishing relationship between 'key values', and that
> structure would usefully be represented in the abstract model?
>
> So descriptions of a high resolution digital image, thumbnail
> and painting might be linked by a common URI reference; a
> description set consisting of 'admin metadata' describing a
> description of a resource, and the description of that
> resource would be related by a common identifier.
>
> I have difficulty understanding how the DCAM can be entirely
> 'ignorant' of such relationships? Elsewhere there are
> attempts to deal with the complexity of such relationships
> eg. FRBR, METS, (MPEG DIDL?)... can the DCAM attempt to group
> descriptions into sets without adressing a more complex data model?
Pete
|