maybe I can clarify some points:
I can not speak for Bibframe, but publishing catalogs as Linked Data is not the primary purpose of Bibframe, this has started much earlier. "Linked Data" is coined as a term for an "easy path" towards Tim Berners-Lee's vision of a Semantic Web that can be understood as an improved version of the World Wide Web of today (a giant global graph with database features)
Bibframe embraces some building blocks of the Semantic Web, the main purpose is to leave all the MARC limitations behind. In the first phase, no dependencies to W3C recommendations or other third party vocabularies should be created inside of Bibframe. The use of W3C recommendations might happen by the community that uses Bibframe, or later by establishing best practice efforts, as far as I understand.
The building blocks like RDF allow catalog data (or better: catalog entities) to be part of the Semantic Web. There will be no more need to "publish" catalogs then, because just exporting library catalog data or exposing them to the web will enable everyone who uses the Semantic Web (by consuming Linked Data) to also use library catalog data too. With MARC, this is obviously not possible.
Personally, I expect rare to no support from web search engine vendors for providing high quality library data. Indexing them into commercial search engines may be an attractive goal for some situations, but it is only a (nice) side effect of a greater project, the opening of library catalogs to the World Wide Web. The web search industry maintains very own strategic projects like schema.org
. So if a library decides to give data to a search engine vendor, they could use schema.org
to express "popular meanings" of entities in a library catalog to those who use the products of the search engine vendors. For example, people searching the web for a "book". There are efforts how to make use of Bibframe for schema.org
vocabulary so this task must not be repeated by each library again and again.