December 19, 2014
Ray Denenberg, Library of Congress
The library/bibliographic community is in the process of developing a model for annotating bibliographic information. This paper attempts to cast the bibliographic annotation model in term of web annotations, and to do so in layman terms (by “layman” we mean someone not familiar with bibliographic information).
To begin, assume LibraryA acquires a (physical) copy of the book Goldfinch (Tartt). It “catalogs” the book, meaning it creates a bibliographic record, or to put it in web terms: “a resource description”, which is an RDF resource. To be clear, the book itself is indeed a resource being described, however the book is not an RDF resource and it is not a resource to be annotated; rather, the resource to be annotated is the resource description. This is a critical distinction in our model.
Cataloging an item (creating a bibliographic record) is an expensive process. So LibraryA doesn’t actually create the record itself; rather, it finds an existing description for the item in a union catalog, and it attaches institution specific information to that record.
Next, LibraryB acquires a copy of Goldfinch (the same publication) and similarly finds the same existing description and it attaches its institution specific information.
Conceptually there are now two RDF resources describing the book Goldfinch:
Much of the information about the book is common to both descriptions, for example, title, author, publisher, date, language, subject(s), ISBN, and so on – information intrinsic to the publication. So ResourceA and ResourceB can be thought of as value-added resources, where the “value” is added to a “pure” description containing intrinsic information only:
So ResourceA is information from ResourceC plus information specific to LibraryA. Similarly, ResourceB is information from ResourceC plus information specific to LibraryB.
The library-specific information is in the form of annotations of resource C .
So, when LibraryA creates its description, the conceptual process is:
And LibraryB goes through a similar process.
In the BIBFRAME view, the purpose of an annotation is to:
Examples of each follow.
There is a NY Times review of Goldfinch at http://www.nytimes.com/2013/10/13/books/review/donna-tartts-goldfinch.html?pagewanted=all&_r=0
An annotation, asserting that this resource is a review of the resource described by ResourceC would look like.
<http://example.com/UnionCatalog/annotations/books/goldfinch/reviewxyz/>
[a oa:Annotation, bf:Review ;
body <http://www.nytimes.com/2013/10/13/books/review/donna-tartts-
goldfinch.html?pagewanted=all&_r=0>
target <http://example.com/UnionCatalog/descriptions/intrinsic/books/goldfinch> .]
There are other reviews:
Annotations similar to the one above may be created for each of these. LibraryA might choose to advertise the NY Times review, while LibraryB might choose to advertise the NPR and Washington Post reviews; both might select the LA Times review, and perhaps neither selects the Washington Times.
A holdings annotation says for example: “This institution holds a copy of this book”. We can look at two different processing models here.
In the first, When LibraryA and LibraryB obtain a copy, they do not necessarily need to create ResourceA and ResourceB as described above. Instead, they can both simply annotate ResourceC, each saying “I have a copy”. A patron, looking for a copy, can retrieve holding annotations for ResourceC and discover that there are copies at LibraryA, LibraryB, and any other library that has attached a holdings annotation.
Alternatively, LibraryA and/or LibraryB may support the capability of a patron to search for holdings other than its own. A patron of LibraryA might want a book that is unavailable at LibraryA (perhaps the library holds a copy but it is out on loan) but LibraryA provides information about nearby libraries holdings of that book. In that case LibraryA would create ResourceA (as described above) which would include holding annotations not only for its own holdings but also for nearby libraries.
“Cover art” refers to an illustration associated with a publication, for example a book. It serves several purposes, one of which is to advertise the book. A book might have several cover art illustrations, and there are many commercial artists supplying cover art. For the example at hand, Goldfinch there are a number of images to choose from, for example:
LibraryA might choose the first and LibraryB, the second, and these choices would be represented as annotations.
