You may want to actually use schema:domain and schema:range rather than rdfs:domain and rdfs:range. Telling people not to infer too much isn't enough because tools may be involved.

Sent from my iPad

On Mar 23, 2013, at 3:25 PM, "Ford, Kevin" <[log in to unmask]> wrote:

Dear Simon,


Jeff's assessment is correct.  At this time, do not infer too much from domain and range with BIBFRAME except to denote expected use.  It is generally safe to assume inheritence from supertypes to subtypes.


Presently, we're striving to be informative without being restrictive.  That's to say that were focused on vocabulary and requirements at this time after which it would be time to explore more formal schema declarations (again, if desirable).







From: Bibliographic Framework Transition Initiative Forum [mailto:[log in to unmask]] On Behalf Of Young,Jeff (OR)
Sent: Thursday, March 21, 2013 9:52 PM
To: [log in to unmask]
Subject: Re: [BIBFRAME] The Bibframe ontology uses multiple domains in a probably unintended way.




I think Kevin is imagining schema:domain and schema:range type meaning as opposed to rdfs:domain and rdfs:range meaning. These are annotation properties that amount to guidelines rather that strong RDFS/OWL assertions capable of sustaining inferences.



Sent from my iPad

On Mar 21, 2013, at 9:41 PM, "Simon Spero" <[log in to unmask]> wrote:

On Thu, Mar 21, 2013 at 4:35 PM, Ford, Kevin <[log in to unmask]> wrote:


Right, your conclusions are correct based on what we've published.  We would have to employ a unionOf axiom to ensure against unwanted collisions.


However, we're interested in minimally constrained semantics and what you are seeing is a nod toward that goal. 


 I think there may be some confusion here.


The current vocabulary is not *minimally* constrained; it is instead closer to being maximally constrained. 


Every domain statement that is added to a property increases the constraints on how that property can be used.  


As things are now, the only things that can be annotated are things that are simultaneously Works, instances, and authorities.  


If you want to be able to annotate some Works but not Authorities, then you have to do something different.  This does not even require using OWL.