On Mon, Jul 22, 2013 at 05:20:46PM -0400, Ray Denenberg wrote:
> There will be no claim of conformance to or compatibility with OAM.
> So in particular:
> > In section 6 of the BF annotation document  there is only one entry
> > under "Subproperty of":
> > bf:annotates -> subPropertyOf -> oa:hasTarget
> This no longer applies.
The loss of this very sensible mapping would be a pity if the basic models --
at the level of what OA calls the Open Annotation Core -- really were
compatible which, as far as I can see, they really are.
> I hope to have more to say on all of this very soon. I will say this
> though: Much of the changes in draft 2 respond directly to the criticisms of
> and confusion over controversial areas of draft 1. In particular, there are
> a few areas where we feel that BIBFRAME Annotations are fundamentally
> different from web annotations large, and there was strong criticism from
> OAM to the effect of "if you are not going to conform to our model, don't
> pretend that you do". So we won't.
When Robert said this , I understood him to mean the full Open Annotation
Data Model , which includes "rules intended to enhance interoperability
between communities" -- _not_ the ontology that defines the properties and
classes used for the data model [3,4,5]. That ontology defines oa:hasTarget
with exactly five triples:
oa:hasTarget rdf:type owl:ObjectProperty
oa:hasTarget rdfs:label "hasTarget"@en
oa:hasTarget rdfs:isDefinedBy oa:
oa:hasTarget rdfs:domain oa:Annotation
oa:hasTarget rdfs:comment "The relationship between oa:Annotation and
target. The target resource is what the oa:hasBody is somewhat "about".
The target may be of any media type, and contain any type of content.
The target SHOULD be identified by HTTP URIs unless they are embedded
within the Annotation. Embedded targets SHOULD be instances of
cnt:ContentAsText and embed their content with cnt:chars. They SHOULD
declare their media type with dc:format, and MAY indicate their
language using dc:language and a RFC-3066 language tag. There is no OA
class provided for "Target" as a target might be a body in a different
annotation. However, there SHOULD be 1 or more content-based classes
associated with the target resources of an Annotation, and the dctypes:
vocabulary is recommended for this purpose, for instance dctypes:Text
to declare textual content." .
The OA documentation could, in my opinion, more clearly draw attention to the
distinction between the OA Data Model and the OA ontology, and it is unusual
for an ontology to include such usage recommendations in an rdfs:comment (see
above), but it is very significant that none of the recommendations say "MUST".
As far as I can tell, the OA and BF annotation approaches apply the same core
model, only they were designed for somewhat different typical use cases so are
associated with different usage guidelines. Unless the meaning of bf:annotates
were changed in ways that REALLY made it incompatible with oa:hasTarget as
defined above, it would be very helpful to keep the mapping. In fact, other
mappings between the vocabularies might helpfully be specified as well.
Doing so would _not_ imply that BF annotations conform with the particular
"rules intended to enhance interoperability between communities" specified in
the OA Data Model.
Tom Baker <[log in to unmask]>