library of congress - wordpress.com · 30-10-2015 · the bibframe editor and the lc pilot the...

NOTSL Fall MeetingOctober 30, 2015

Library of CongressBIBFRAME Pilot

Presenter

Presentation Notes

The Library of Congress (LC) launched a BIBFRAME Pilot in June 2015. The pilot commenced with training in Linked Data and Semantic Web principles, and in September 2015, forty Library of Congress catalogers began creating BIBFRAME descriptions for monograph, serial, notated music, sound recording, audio visual, and cartographic resources. This presentation will discuss the background to the LC BIBFRAME Pilot – the training in Linked Data and Semantic Web principles, the development of BIBFRAME Editor RDA profiles, and the hands-on instruction in using the BIBFRAME Editor RDA profiles. Successes as well as bumps along the way will be covered.

THE BIBFRAME EDITOR AND THE LC PILOT

The Semantic Web and Linked Data : a Recap of the Key Concepts

Presenter

Presentation Notes

Welcome to “The BIBFRAME Editor and the LC Pilot” -- the third module developed for LC catalogers who will be participating in the LC BIBFRAME pilot.

Learning Objectives• Describe the goals and some key concepts of the Semantic Web and Linked Data

• Explain libraries’ role in a Linked Data environment

• Iterate some of the major aspects of Resource Description Framework (RDF)

• Discuss the basic concepts of BIBFRAME

Presenter

Presentation Notes

Semantic Web, Linked Data,and Libraries

• “What are the goals and purposes of applying the principles of Semantic Web and Linked Data to library data?“

• The Why? of “Why? What? Who?”

BIBFRAME: Why? What? Who? (Word : 628 KB) (May 1, 2014)

Presenter

Presentation Notes

http://www.loc.gov/aba/pcc/bibframe/BIBFRAME paper 20140501.docx

Some Answers to the Question• Increase the visibility and usage of Library data on the Web

• Integrate library data with the large number of structured data sources and links on the web

• Create relations among resources• Enhance the sharing of library data with a wider audience

• Facilitate a more full implementation of RDA

• “A virtual stack browsing experience”

Presenter

Presentation Notes

The Most Practical Goal?

• Linked Data is to replace MARC as the primary carrier of library data– a standard machine readable format– using common web standards– MARC is used chiefly by libraries, and not as well understood by other communities

• Transition: from a static two‐dimensional collocated record to decentralized data with links to illuminate relationships

Presenter

Presentation Notes

MARC Record

Humans can ‘connect the dots,’ but …

Presenter

Presentation Notes

[‘Humans’ text box appears on 1-second delay]

RDF Graph of a MARC Record

3‐8September 2015

Presenter

Presentation Notes

What you see on this slide is a small portion of the RDF graph for the MARC record on the previous slide. Such a graph enables computers to ‘connect the dots.’ The record was downloaded from OCLC as a MARC RDF/SML file, converted into a BIBFRAME record using MarcEdit, opened in Notepad ++, copied, and parsed using an RDF validator which produced the graph. But don’t let this level of detail intimidate you!

The Four Principles of Linked Data

• Use URIs as names for things – “identifiers”– Not data strings, which computers can’t interpret

• Use HTTP URIs so people can look up the names– e.g., id.loc.gov

• When someone looks up a URI, provide useful information, using standards– e.g., Resource Description Framework (RDF); SPARQL

• Include links to other URIs, so humans and computers can discover more things

Presenter

Presentation Notes

Tim Berners-Lee outlined these four ‘rules’ of Linked Data.�

Libraries’ Important Role in aLinked Data World

• No other community works with authorities as much ‐‐ or as well ‐‐ as libraries do

• Libraries are experienced at identifying, structuring, and organizing data in many ways

• But …– We need to translate MARC skills and practices into a Linked Data context

Transition from authorities to identifiers

Presenter

Presentation Notes

Are identifiers and authorities more-or-less the same thing? Two ways of looking at the same concept? Identifiers are more easily maintained than data strings of ‘authority records’.

Linked Data’s Benefits for Libraries

• Wider accessibility of library data• Cost savings through increased data sharing• Transition beyond proprietary MARC format• Reduced duplicative processing• Less time spent on authority maintenance, through use of identifiers rather than text strings– Language neutral

Linked Data 5‐star Scheme(Tim Berners‐Lee)

**********

*****

• Make your stuff available on the Web • Make it available as structured data• Use non‐proprietary formats• Use URIs to identify things, so that people can point at your stuff

• Link your data to other data to provide context

http://www.w3.org/DesignIssues/LinkedData.html

Presenter

Presentation Notes

Tim Berners-Lee introduced a 5-star rating to encourage data publishers along the road to good Linked Data. “Under the star scheme, you get one (big!) star if the information has been made public at all, even if it is a photo of a scan of a fax of a table. You get more stars as you make it progressively more powerful, easier for people to use.” Source: “Is Your Linked Open Data 5 Star?”

http://www.w3.org/DesignIssues/LinkedData.html

RDF: Resource Description Framework

• Standard model for exchange of data on the Web

• Structures relationships between resources, people, and things on the web

• Uses graph model to represent database relationships

• RDF and related standards maintained by the World Wide Web Consortium (W3C)

Relax!• You won’t need to know all the details of RDF in order to use the BIBFRAME Editor– Certainly not to the extent you needed to learn MARC! (i.e., by endless rote memorization)

• In the past, a deep knowledge of the container (MARC 21) was needed for cataloging

• With BIBFRAME, cataloger input is needed for the development of profiles for the Editor

Presenter

Presentation Notes

Last bullet: The course developers (former catalogers) have been working hard with the programmers to achieve this.

BIBFRAME Editor

• Profiles: basis for entering data in the Editor– Under development for the Pilot– Mapping to RDA elements– Mapped by format

• BIBCO Standard Record (BSR)• CONSER Standard Record (CSR)

• Look‐up capability for authorities and bibliographic descriptions

• Meaningful RDA captions in the interface

• What follows is intended to help you better understand the broader context within which the Editor functions

RDF Tools• URIs: to identify resources and relationships• Vocabularies and ontologies: tools that define relationships between resources

• Triple statements: the core means of expressing relationships

• Standard languages: to express relationships• Query languages: allow people and machines to interact with RDF data in large data setsCommon usage of these tools promotes wide use and reuse of Linked Data on the web

Presenter

Presentation Notes

Important to note that we are talking about Uniform Resource Identifiers not URL “locations.” URIs are used to identify resources and relationships -- easy for machines to point to a specific meaning associated with a URI (strings or “literal” values are more difficult for a machine to parse). Often you will see the term IRI (Internationalized Resource Identifier) which refers to a type of URI. For the purposes of this workshop we will refer to URIs and IRIs as one in the same. The terms ‘vocabularies’ and ‘ontologies’ are used somewhat interchangeably in this presentation. They both refer to a way to define concepts and relationships within a particular domain (such as the LCSH vocabulary). The trend is to use the word “ontology” for more complex, and possibly quite formal collection of terms, whereas “vocabulary” is used when such strict formalism is not necessarily used or only in a very loose sense. Vocabularies are the basic building blocks for inference techniques on the Semantic Web.

RDF Data Model

• Triple statements• RDF XML

– or other serialization formats (e.g., Turtle, N‐triples)

• URIs• Namespaces• Ontologies and vocabularies

Triple Statements

Subject ObjectPredicate

“This work” “This author”“Was written by”

• Two things, and a relationships between them

Presenter

Presentation Notes

The purpose of a triple is to make statements about resources. The statements can be represented in a graph as shown on this slide. Graph data is a model that simply refers to statements about relationships between resources, books, people … almost anything that can be identified and represented in a way that computers can process the information about the resources and relationships The subject, predicate, and object are the basis of the triple statement and are identified in this graph model of the statement: “This work was written by this author.”

Triple Statements• Subject: identifies a “Resource of interest”• Predicate: identifies a Property of the “resource of interest” ‐‐ a relationship

• Object: identifies a Property value ‐‐ a resource that has a relationship to the “resource of interest”

• Ideally, all three are identified by URIs

Triple Statements

This land is your land

Woody Guthrie

Was written by

URI forwork

URI for author

URI for Dublin Core term: Creator[read: has creator]

Presenter

Presentation Notes

The top example shows an English-language triple statement of the relationship between a book and its author. The bottom example shows what a machine needs to digest the statement and serve meaningful information to a user. In a minute , we’ll see what that triple looks like when the machine has what it needs to do that. (Photo credit: Library of Congress)

Triple Statements

The triple statement: This land is your land has creator Woody Guthrie

Can be expressed in a way machines can interpret using URIs for name authorities and for Dublin Core terms:

<http://id.loc.gov/authorities/names/n2013032388><http://purl.org/dc/terms/creator><http://id.loc.gov/authorities/names/n79111488>

Presenter

Presentation Notes

Note that the predicate in this triple uses the Dublin Core (i.e., “dc”) term for the property/relationship: “has creator”. Next, we will see the ‘graph’ view of this triple.

Triple Statements ‐‐ The Next StepThe Graph View

Subject Predicate Object

This land is your land has creator Guthrie, Woody, 1912‐1967

* http://validator.w3.org/

Presenter

Presentation Notes

And this is the graph view of that triple, as run through a ‘W3C markup validator’

Triple Statements – The Crux

• Triple statements make it possible to make meaningful statements about resources on the semantic web

• Can be processed by computers and serve meaningful results to users

RDF XML

• Uses XML structure to help computers read statements about resources

• Format for expressing triples– Identifies the syntaxes and vocabularies used to express triple statements

– URIs used to identify resources and namespaces– Namespaces identify the vocabularies and syntaxes used to make semantic statements about resources

RDF XML – ‘Under The Hood’ <?xml version="1.0" encoding="UTF‐8"?><rdf:RDFxmlns:rdf="http://www.w3.org/1999/02/22‐rdf‐syntax‐ns#"xmlns:dc="http://purl.org/dc/elements/1.1/"xmlns:lcnaf="http://id.loc.gov/authorities/names"><rdf:Description rdf:about="http://id.loc.gov/authorities/names/n201303238><dc:creator>http://id.loc.gov/authorities/names/n79111488</dc:creator></rdf:Description></rdf:RDF>

Presenter

Presentation Notes

Here is our triple statement ‘This land is your land has creator Woody Guthrie’ expressed in RDF XML On the next few slides, we will ‘take-apart’ this file.

RDF XML – ‘Under The Hood’:Interpreting the File

<?xml version="1.0" encoding="UTF‐8"?> Document is XML

<rdf:RDF and </rdf:RDF> The Root, or “wrapper”, of all the contents of the file

xmlns:rdf="http://www.w3.org/1999/02/22‐rdf‐syntax‐ns#“Namespace: identifies RDF as the syntax used

xmlns:dc="http://purl.org/dc/elements/1.1/“ Namespace: identifies Dublin Core as source (of the term used in predicate)

xmlns:lcnaf="http://id.loc.gov/authorities/names"> Namespace: identifies the LC NAF as ID (of subject and object)

Presenter

Presentation Notes

The meaning of each statement is provided underneath, in bold.

RDF XML – ‘Under The Hood’:Interpreting the File

<rdf:Description and </rdf:Description>Beginning of triple End of triple

rdf:about="http://id.loc.gov/authorities/names/n2013032388“ Subject

<dc:creator>http://id.loc.gov/authorities/names/n79111488</dc:creator> Predicate and Object

(this presentation is a valid RDF XML shortcut)

Presenter

Presentation Notes

In the last two lines, the predicate and object are reversed from the order which we have seen until now. This is a valid RDF XML shortcut that machines are happy to digest. In this display, a single line of code contains the reference to the predicate (beginning with: <dc:creator>) that encloses the reference to the object.

URIs in RDF XML

• Used to retrieve content to be read by both humans and machines

• Two forms:– Humans get an HTML page to read– Machines retrieve an RDF XML format (or another format) that it can interpret and act on

The Two Forms

Presenter

Presentation Notes

Here we see both forms for another entity., made possible by the use of RDF XML. In the background is the ‘human’ form. This is from VIAF, but it is not that different from the LC NAF that you are used to viewing with your and eyes and interpreting with your mind. Superimposed on top is the ‘machine’ form

URIs in RDF XML

• URIs identify web resources – Book or author– Namespaces of standards that have been used to encode triple statements

– Vocabulary and ontology terms– Triple components

• Subject• Predicate• Object

Presenter

Presentation Notes

Note: Objects in a triple statement can be literals, IRIs or blank nodes

Namespaces

• Declared in the root of an XML file• Identified by URIs• Declare:

– Vocabularies– Syntaxes– Sources of terms used to describe and identify the resource

Namespaces – Examples

• xmlns:rdf=http://www.w3.org/1999/02/22‐rdf‐syntax‐ns#

• xmlns:dc=http://purl.org/dc/elements/1.1/• xmlns:lcnaf=http://id.loc.gov/authorities/names

Presenter

Presentation Notes

These are the specific namespaces we saw in our ‘under the hood’ example

http://www.w3.org/1999/02/22-rdf-syntax-ns

http://www.w3.org/1999/02/22-rdf-syntax-ns

http://purl.org/dc/elements/1.1/

http://id.loc.gov/authorities/names

Vocabularies and Ontologies

• Used to define concepts within a particular field of study (domain)

• Define classes of objects• Define properties of resources• Define relationships between objects• Can be expressed using RDF, so computers may interpret them

• Help retrieve meaningful search results

Presenter

Presentation Notes

Ontologies are a formal way to describe taxonomies and classification networks, essentially defining the structure of knowledge for various domains -- the nouns representing classes of objects and the verbs representing relations between the objects.

Vocabularies and Ontologies –Example of Discovering Relationships

• Data set says “Flipper is a dolphin”• Ontology says “all dolphins are mammals”

• A semantic web program that understands that X = Y …

• Can discover a new relationship: “Flipper is a mammal”

Presenter

Presentation Notes

Discovering relationships is also called inference, see http://www.w3.org/standards/semanticweb/inference The data set to be considered may include the relationship (Flipper is a Dolphin). An ontology may declare that “every Dolphin is also a Mammal”. That means that a Semantic Web program understanding the notion of “X is also Y” can add the statement (Flipper is a Mammal) to the set of relationships, although that was not part of the original data. One can also say that the new relationship was “discovered”.

BIBFRAME is Only One RDF Vocabulary

• Others– Dublin Core– FOAF (‘Friend of a Friend’)– Library of Congress authorities and vocabularies at http://id.loc.gov

– RDA vocabularies and registry: http://www.rdaregistry.info/

– Schema.org• All have namespaces• All define classes, elements, and values

Presenter

Presentation Notes

The RDA Registry contains vocabularies that represent the RDA Element set, relationship designators, and controlled terminologies as RDA element sets and RDA value vocabularies in Resource Description Framework (RDF). RDA element vocabularies are available in published form in the namespace

http://id.loc.gov/

http://www.rdaregistry.info/

http://www.schema.org/

BIBFRAME Vocabulary• Work ‐ reflects a conceptual essence of the

resource• Instance ‐ reflects an individual, material

embodiment of the Work• Authority ‐ defined relationships reflected in

the Work and Instance– People, Places, Topics, Organizations, etc.

• Annotation ‐ enhances our knowledge aboutanother resource– Library Holdings, Cover Art, Reviews, etc.

BIBFRAME/RDA ‐‐ Terminology

BIBFRAME Work

BIBFRAME Instance

BIBFRAME Annotation

RDAWork

RDA Expression

RDA Manifestation

RDAItem

Presenter

Presentation Notes

Simply put, this is how the main BIBFRAME and RDA terms relate to each other – it’s not that hard to learn the minor differences.

BIBFRAME Classes

Source: http://www.loc.gov/bibframe/

http://www.loc.gov/bibframe/

BIBFRAME Properties

And this is only some of them … !

Presenter

Presentation Notes

(From the same source as the preceding slide)

Now you’re conceptually ready to start working with the BIBFRAME Editor

• Remember that you are participating in a “pilot”• Remember that the Editor is a work‐in‐progress• Remember that this is an entirely new way of displaying and looking at bibliographic data– It is not necessary that BIBFRAME do all that MARC did

• During the Pilot, we want you to …– identify what works and what doesn’t– suggest enhancements– provide constructive feedback– be flexible and not stress over what is ‘uncomfortable’

Presenter

Presentation Notes

Linked data and free-floating subdivisions are not compatible!

library of congress - wordpress.com · 30-10-2015 · the bibframe editor and the lc pilot the...

Documents