On Wed, Jun 25, 2014 at 12:29 PM, J. McRee Elrod <[log in to unmask]> wrote:
My concerns are several.  These are my three primary ones:

Concern 1:  

Theoretically, I am concerned with the move from language neutral numbers to ambiguous English words as element identifiers.  Remember the Dublin Core user who entered the person who gave the item to the
library under "Contributor"?  This is of particular concern to us in a bilingual country, and with multilingual overseas clients.

Semantics cannot depend on the lexical form of a label; this is a well known  anti-pattern in knowledge representation systems (see e.g. Drew McDermott's famous essay "Artificial Intelligence meets Natural Stupidity").    

Fortunately, the IRI used to identify a concept is distinct from the label be used to present it to one viewing the ontology.  OWL, RDF(S), and SKOS allow for different textual labels, in many different languages and dialects, to be applied to a class or property. 

Using E78_Collection from the CIDOC CRM as an example, we can see that the class has labels in six different language.  That's four louder than bi-lingual, and six more than natural language terms than some random number like 880. 

You can also see that the labels do not convey the meaning;  that arises from the text of the CRM, as located in the comment.    

        rdfs:comment "This class comprises aggregations of instances of E18 Physical Thing that are assembled and maintained (“curated” and “preserved,” in museological terminology) by one or more instances of E39 Actor over time for a specific purpose and audience, and according to a particular collection development plan.  
Items may be added or removed from an E78 Collection in pursuit of this plan. This class should not be confused with the E39 Actor maintaining the E78 Collection often referred to with the name of the E78 Collection (e.g. “The Wallace Collection decided…”).
Collective objects in the general sense, like a tomb full of gifts, a folder with stamps or a set of chessmen, should be documented as instances of E19 Physical Object, and not as instances of E78 Collection. This is because they form wholes either because they are physically bound together or because they are kept together for their functionality.
        rdfs:label "Sammlung"@de,
        rdfs:label "Collection"@en,
        rdfs:label "Collection"@fr,
        rdfs:label "Συλλογή"@el,
        rdfs:label "Коллекция"@ru,
        rdfs:label "Coleção"@pt
Concern 1- alleviated. 

 Concern 2:
Pragmatic ally, I am concerned with the investment we have in MARC based programs and tools.  We offer our clients an AACR2 compatible export of MARC/RDA records, as well as a UKMARC export for British
libraries still using UKMARC (fewer and fewer).  Perhaps we will explore a Bibframe export, but continue to use unambiguous MARC number labels and our MARC based systems for record production?

As a dedicated, for profit MARC record provider, this is a valid concern.  

Concern 3: 
Financially, a move to Bibframe would create an even greater divide between have and have not libraries, than did the RDA Toolkit.  We had one small client which RDA caused to abandon their catalogue, and move
to spreadsheets.  We suspect many small libraries will key from CIP into their present systems (probably minus fixed fields), rather than attempt to make the move.

This is a strong claim, that if true, casts serious doubts upon the design of BIBFRAME.  Assuming arguendo that this does hold, it does not entail that a different Ontology of the Bibliographic Universe would have the same effects.

The use of Linked Open Data to provide access to bibliographic information makes this information more available than if it is only available by purchase or from manual creation.  A properly designed infrastructure should thus be making more information available to poorer libraries.  

Since the hypothesis posits the presence of CIP information, it is not beyond the bounds of reason to expect that the items being inventoried may have identifying numbers associated with them, which can be used to locate descriptive information, without rekeying.  Unique items requiring original cataloging tend to be the preserve of the wealthy, who are pre-adapting, or of the archivist, for whom MARC has at best been a mixed bag of materials.