I think it's c) except that the unknown statements don't necessarily need to be "dropped". They only need to be "carried" until such time as the software catches up.

The idea of "linked records" is odd. "Linked things" is more like it. The difference is that deployed systems may not understand certain statements (datatypeproperty or objectproperty) that detail those things.

Jeff

Sent using OWA for iPad
From: Bibliographic Framework Transition Initiative Forum <[log in to unmask]> on behalf of Stuart Yeates <[log in to unmask]>
Sent: Wednesday, June 29, 2016 8:11:39 PM
To: [log in to unmask]
Subject: Re: [BIBFRAME] FW: [BIBFRAME] rdf:value
 
Take the example I mentioned, a library consortia. If new predicates are added to the namespace and those predicates are shared by one consortia member before other consortia members have updated themselves. Two things can then happen either:

(a) sharing errors out / fails
(b) the importing library has to update its definition of the entire BIBFRAME vocabulary
(c) predicates are dropped

Going with (a) breaks the functioning of the consortia.

Going with (b) puts us in the situation where any import can force an BIBFRAME compliant system to completely rebuild itself around a new definition of BIBFRAME which could take an arbitrary length of time, which is not really something we want except during scheduled outages. Caveats around 'minor changes' don't work, because that's exactly the stuff on which the halting problem is built. Note also that in my original post I used the expression "in the general case".

Going with (c) puts us in the situation were we end up with potentially significant differences in the data across the consortia. If the predicates are used in bibliographic control or linking, those differences could cascade across linked records.

cheers
stuart
--
I have a new phone number: 04 463 5692
https://www.facebook.com/VUWLibrary / https://www.facebook.com/TKMPC

________________________________________
From: Bibliographic Framework Transition Initiative Forum <[log in to unmask]> on behalf of Simon Spero <[log in to unmask]>
Sent: Thursday, 30 June 2016 9:08:29 a.m.
To: [log in to unmask]
Subject: Re: [BIBFRAME] FW: [BIBFRAME] rdf:value

On Jun 29, 2016 4:40 PM, "Young,Jeff (OR)" <[log in to unmask]<mailto:[log in to unmask]>> wrote:
>
> I donít think this is true. A vocabulary should publish versions at new URIs but the namespace can remain constant.

Stronger than that - The latest version of an ontology can be published at a fixed IRI, as long as each version has a unique version IRI.

In fact there is no necessary relationship between an ontology, any names mentioned in that ontology, and any URLs that can be used to retrieve that document.

There are some recommendations as to how ontology IRIs, ontology version IRIs, and URLs should behave for ontology version series, but these are advisory, and behavior is under constrained.

The only time that new names are needed is if the semantics of the previous name is changed. *Cough* skos:broader.

Simon