Dear all,
[first of all: Thank you very much for your detailed comments (part 1), Pete!]
[snip]
> > 1) type of collection
> >
> > We would like to convey two kinds of information:
> >
> > 1. This IS a collection of (retro-)digitized material
> > (all of our collections are of that same type).
> > 2. This is a collection of certain item types (newspapers,
> > encyclopedias, pamphlets, journals).
> >
> > At first glance, cld:CLDType seems to be appropriate
> > for 1. But then we would have to use dc:type
> > cldtype:collectionText, and that's too generic for
> > our purpose. We would like to be able to distinguish
> > collections of "born digital texts" from those containing
> > "digitized texts". I believe Issue 2 would be covered by
> > cld:itemType I think.
>
>Just to clarify, the property cld:itemType was being suggested as an
>alternative approach to the current use of dc:type i.e. rather than
>saying
>
>collection:C dc:type colltype:CollectionText . ("Collection C is a
>member of the class of collections with items of type Text")
>
>we might say
>
>collection:C cld:itemType dcmitype:Text . ("Collection C has items of
>type Text")
What do you think of this example:
dc:type cldtype:CollectionText
dc:type pddtype:CollectionDigitizedText
cld:itemType pdd-pubType:encyclopedia
>In both cases, I think it would be possible to define a "local" type
>vocabulary to specify a subtypes/subclasses of digitised textual items
>or "born digital" textual items
>[snip]
> > 2) institution responsible
> >
> > Should be dc:creator.
[snip]
>Note that the DC CD AP uses the property dc:creator to indicate the
>"collector" of a collection. That may or may not be the agent that
>carried out the digitization (because I imagine those digital items
>might be grouped into other collections after their digitization) - but
>in your case it sounds like the collector is the digitizer. And, yes, as
>you say, the collector of the physical collection and the collector of
>the digital collection are probably different agents.
The important information in our case is: "Who created the digital
collection?" (Which often means: who is responsible, who administers the
resource, who offers the service ...). Since I cannot see another
appropriate element to carry this informationen, I would use dc:creator. I
understand that this differs from the DC CD AP definition and would welcome
any suggestion how to convey the described information in a better way
(closer to DC CD AP).
> > 3) location
> >
> > gen:isLocatedAt. Which location is meant: Location of the
> > physical collection? Location of the server? Location of the
> > institution which hosts the (digital) collection? Location
> > of the institution which created (cf: dc:creator) the
> > (digital) collection?
>
>We still have an outstanding task (the major task outstanding, I think)
>to clarify the distinction between the Location and Service entities and
>the isLocatedIn and isAccessedVia relations/properties.
Consequently here I would have to give an information on the location of
the creator of the digital resource. Although users might find it more
interesting where the physical items in the collection come from ...
[snip]
> > 4) cooperation partners
>
>Again, it depends exactly what this means. What is the relationship
>between the collection and this agent/these agents?
[snip]
>AFAIK, so far no-one has expressed a requirement to represent some
>information that we would use the dc:contributor property for. But if
>there is a requirement that isn't covered, then we can consider whether
>we need to include it (and how it should be represented).
>
>Could you maybe say a bit more about what relationship you want to
>represent please?
I think we have at least five types of contribution or cooperation in our
case of digitization-projects:
- financial support (funding)
- technical support/services
- additional delivery of data / metadata
- provision of the physical items to be digitized
- participation of additional institutions in the project (in case
of a wider project-consortium for example)
I do see the problems with xml-attributes. But I don't see a way to work
around this by a controlled vocabulary for example.
>[snip]
Best, Patrick
|