1. Looking to the Future with RDA
This presentation is based on earlier presentations of the same title given during the American Library Association Conference and other meetings internationally since 2009 and recorded for the Library of Congress in It still applies and gives you my personal views for the future and how RDA fits into all of this. These are my own views and do not necessarily represent the views of the Library of Congress.
It’s really about the need to develop better information systems for the future to better serve our users.
RDA: Resource Description and Access is designed as a content standard for the digital environment, but will also function to create bibliographic descriptions and authority data for any cataloging scenario – whether for book or card catalogs or current integrated library systems or systems of the future. It’s based on the conceptual models of FRBR (Functional Requirements for Bibliographic Records) and FRAD (Functional Requirements for Authority Data) and on the IFLA International Cataloging principles.
Presented by Dr. Barbara B. Tillett
Chief, Policy & Standards Division, Library of Congress
For National Central Library and
Library Association of the Republic of China (Taiwan)
March 21, 2012
Library of Congress RDA Seminar, March 2012 1 1

2. “…we’re now focused on three things: ease of use, continuous availability, and low price.” Robert Capps in the Sept Wired article “The Good Enuf Rvlutn”
“…multiple, simultaneous ways of organizing things” – David Weinberger, Everything is Miscellaneous, 2009.
Since 2009 we’ve seen more and more writers about the Web environment and access to information, telling us of the importance to users of making it easy to find what they want, of having systems that are always available and at a low price (mostly free), but also reminding us that our old ways of trying to provide just a single way to organize things as we have with catalogs and authorized access points, isn’t the only way to look at the bibliographic universe.
We can have complementary ways of organizing things to open up more pathways for users to find what we have in our library collections and related resources beyond our libraries.
That point of view is now, and yet our libraries and catalogs are still operating as if we are in the 1960’s.

3. Library catalogs of the past were stand-alone end points to what was in a particular collection, and typically had to be used within the walls of the library, unless you had a book catalog that you could buy or borrow from someone.
Libraries have made great strides toward a Web presence, but many offer only an electronic version of their card catalogs. Linear displays of citations to holdings may include a link to a digitized version of the described resource, but typically excludes machine-actionable connections. Citation-based catalogs need to describe resources by their identifying characteristics in a way that computer systems can understand and show relationships to persons, families, corporate bodies, and other resources. This will enable users to navigate through linked surrogates of resources to get information more quickly. It also will better enable systems to make cataloging easier. 3

4. Internet
Information systems are no longer the end points in isolation of old catalogs
Global access to data
Integrate bibliographic data with wider Internet environment
Share data beyond institutions
We then moved on to at least putting those old catalogs onto the Web so they could be searched, and OCLC even took it one step further to do a WorldCat to offer users a way to search the holdings of all the OCLC member libraries at once. But in fact all of that bibliographic data from any source <click> can now be integrated into the wider Internet environment. New kinds of links can be made, new displays can be generated for users from data packaged in new ways – all of it on a global scale in multiple languages and scripts. We now have the technology and are providing global connection anywhere that computers can operate – that includes the digital connections of cell phones and smart phones with Internet access.
The computer systems know where the user is located based on the GPS (global positioning system) and they can use that information to suggest nearby libraries that have the resources they select to use. OCLC’s WorldCat for a several years now already has a variation of this location specific feedback to users with Google’s “Find in a Library” feature. But they can also offer where a user can purchase a resource they want or view or listen to a digital version of information they want.
The new cataloging code, RDA: Resource Description and Access, is designed to help us transition to the technological capabilities of the Internet, today and into the future by having us identify the entities and relationships at the level that machines can use better than they have been able to in the past in our MARC records. However, RDA will also work when we package the elements in MARC records as we will have to do for some transitional period. RDA is not an encoding system or a presentation standard for displays, but instead specifies how to describe the things in our bibliographic universe - resources, persons, corporate bodies, etc., and the relationships among those things. 4 4 4 4

5. Internet “Cloud” Services Databases, Repositories Web front end
The information systems and content – that is the information itself - may all be freely accessible on the Web or available for some nominal fee. I imagine it as something like the Internet cloud computing that we have today with Amazon, Google, Yahoo, and other systems where the elements that describe our resources are available to libraries and users everywhere in the world – not just on an institution’s computer, but shared and available to everyone through Internet search capabilities. The data will come from publishers, from the creators of the resources, from trusted libraries and other institutions, and be augmented by further descriptive information from anyone who wants to help. Some of this is being done now. All of the information about resources in the bibliographic universe will be accessible by any user anywhere at anytime.
Bibliographic data and digital resources are on the web now and we’ve started adding the controlled vocabularies to help identify resources – such as the controlled values for naming the types of content (like sound, text, still images, and so on), types of carriers (like a film reel, a computer disc, a volume), and other elements in RDA that have controlled lists of values- they are already being registered on the Web and can be used to present displays and show pathways to related resources. 5

6. Infrastructure to Build for the Future
Delivery to users
Information Systems + Content
I think we’re at a crucial time for the development of new information systems, more global in nature, more Internet oriented, that can make cataloging easier and make the results of cataloging much more flexible and useful to our users and perhaps more importantly, that enhance data that comes with resources collected by libraries to make them more findable and accessible to our users in the global, Web environment.
RDA is taking the first steps in the right direction. We are building the infrastructure, preparing the building blocks to get there. We can expect a gradual change over several years with adjustments to RDA until there is consensus on this common purpose and we are able to migrate legacy data and build future complementary data that is principle-based and better formed for re-use by search engines and computer systems of the future. Standing still with AACR2 and other old cataloguing codes is not an option if libraries are to remain viable. We knew this in 1997, when the Joint Steering Committee (JSC) held an international conference on the future directions for AACR in Toronto, Canada. That meeting set in motion the work to build the foundations needed for future resource discovery systems. <click> We have the FRBR and FRAD conceptual models with user tasks, <click> we have the International Cataloging Principles from IFLA giving us principles and objectives and some starting rules, and <click> we have RDA building on those foundations to give guidance on specific elements needed for identifying entities and relating them; RDA even includes some lists of controlled terms to use for identifying some elements. For now we have the MARC format, Dublin Core, MODS and MADS and some other schemas to package the data and we have crosswalks among the various communication schema, but we do not yet have agreed <click> data models or <click> new systems to help libraries contribute our resource descriptions to the global Web community to reach the maximum benefits from these changes to <click> deliver information to our users in new ways.
You do not need to know all of the details of the FRBR conceptual model, just as you don’t need to know the electronics behind using your cell phone, but it is helpful to know some basics and to understand the terminology so we as catalogers, as librarians, can work with system designers and search engine companies to build future systems to create and mine metadata to connect our users to our collections. FRBR and FRAD models describe the relationships and connections in our bibliographic universe that in turn can be used to design systems that will enable users to navigate through this universe to things they need or may like to know about.
Data Models
Content Standards
Objectives
Cataloging Principles
Conceptual Models
User Tasks 6

7. 1970s-now MARC structure Self-contained records
May or may not have any explicit connection between bibliographic and authority records
Author/title/subject
Authority record
Bibliographic record
For now, most of us are in an environment that we’ve been in since the 1970’s or late 1960’s when MARC first appeared. We are using the MARC format in an integrated library system of self-contained records.
These records may or may not have any connection to each other and are typically displayed as text in linear displays organized by date entered into the system, or by publication date or by author or title or subject.
I personally hope that RDA will inspire us to move beyond this way of thinking to work with systems designers to develop much better systems for the future that are linked and FRBR-based , showing pathways to related resources. There are some current implementations of FRBR, like OCLC’s WorldCat, VTLS’s Virtua system, Ex Libris’s Next Generation ILS, and experiments at the National Library of Australia and National Library of Sweden, as well as Variations 3, the music catalog project at Indiana University, OLAC’s (Online Audio-visual catalogers) FRBR-inspired prototype audio-visual “discovery interface”, and Kent State University’s research about FRBR - just to name a few, but we still need more, and I hope RDA will give the impetus to move forward.
As I mentioned before, between now and a linked data scenario, we will go through a transition or bridge period.
Holdings/Item record
Holdings/Item record 7 7 7 7 7

8. “Bridge” Period
Mapping tables for RDA and MARC, Dublin Core, MODS/MADS, and ISBD
Decisions from PCC and local choices for alternatives and options, if needed – documented in RDA Toolkit
Development of RDA Workflows and other training materials (e.g., changes from AACR2)
We already have some tools in RDA to help during this bridge period: There are tables that show the crosswalks or mapping between the RDA elements needed for identifying each entity and the corresponding tags or elements in MARC and ISBD (the International Standard for Bibliographic Description), and there are plans to also include such mappings for Dublin Core, MODS and MADS, and other schema, like RDF XML. So, we can package RDA descriptions in the currently used schemas – at least clearly labeled so we can eventually move on. RDA itself is in a bridge period where we have had to carry forward a lot of past practices in order to reach initial agreement for the first release of the instructions. <click>
For the US Test of RDA, the Policy & Standards Division at the Library of Congress has documented the decisions for the Library of Congress regarding our choices for RDA’s alternative instructions and options – these appear in RDA along with the instruction that they address as the LC Policy Statements, or LCPS. We updated those decisions based on findings from the test.
To help current catalogers, there are features of RDA, such as the “Workflows” <click> for simple step by step processes for catalogers to build records. And there also are training materials prepared in the United States, and Webcast training modules and handouts with examples and exercises that are posted on the Library of Congress Web site – I have the URL at the end of this presentation. 8 8

9. Linked Data Scenario Linked “description sets” for entities
Concept
Person
Linked “description sets” for entities
works, expressions, manifestations, items, persons, corporate bodies, families, concepts, etc.
Work
Person
Expression
Manifestation
RDA, by focusing on the entities we want to describe and the descriptive elements that identify those entities, is meant to change the way we think about bibliographic description and access to help us move into the future of linked data and the semantic web – to make our data available for many creative uses on the Web.
RDA has us building linked clusters of data describing each of the FRBR entities, making the relationships explicit. All of this data can be mined and displayed in different ways depending on the user task or application.
Item
Manifestation
Corporate body
Item
Item 9 9 9 9 9

10. FRBR-Based Collocation
Display
All the works associated with a person, etc.
All the expressions of the same work
All the manifestations of the same expression
All items/copies of the same manifestation
Shakespeare
Hamlet
Romeo and
Juliet
English
French
German
Spanish
I hope future systems will be developed to take full advantage of mining the metadata that catalogers provide and have been providing for centuries. It should be easier to fulfill the functions of a catalog to display all the <click> works associated with a person, all the <click> expressions of the same work, all the <click> manifestations of the same expression, and all the <click> items and their special characteristics,
México City
2008
Library of Congress
Copy 1
Green leather binding 10 10 10 10 10

11. Related Works FRBR-Based Collocation
Derivative
works
Stoppard
Rosencrantz & Guildenstern
Are Dead
Movies …
Shakespeare
Hamlet
Subject
Romeo and
Juliet
English
French
Text
German
Spanish
plus all related works: <click> to movies or plays based on Hamlet – all of this to guide a user through our rich collections and beyond for example <click> this shows the connection to the Wikipedia article about Hamlet (a subject relationship) –or we could connect to other related Web resources. It’s not just a hierarchy, but a whole network of related things we have to offer users. And there is so much more data there that could be mined.
Once we are able to share this linked data on the Internet, we can contribute to resource discovery systems and do it through software that will make cataloging much easier by describing once the works/expressions and easily linking new manifestation data to that work/expression data for new resources. When a cataloger begins the process describing a new addition to the already organized body of information in our library collections, our cataloging systems should be able to easily display the associated names, titles, and other data that identify the resource they are cataloging – bringing in data that may come with the resource, such as we do now with ONIX data from publishers, and show the cataloger existing relationships and let them easily establish new relationships for material they are cataloging. For those works that are translated or issued in new editions over time, the systems should let the cataloger easily link to the work/expression level data that already exists – anywhere in the world, and also link names and subjects in our controlled vocabularies.
México City
2008
Library of Congress
Copy 1
Green leather binding 11 11 11 11 11

12. Data Value Registries for Controlled Vocabularies
Values for attributes, e.g.:
Work-Expression content types (RDA/ONIX)
Manifestation carrier types: (RDA/ONIX)
Concepts: (e.g., LCSH subject heading strings)
Categories of entities
FRBR & FRAD: work, expression, manifestation, item, person, family, corporate body, concept, object, event, place
Registries on the Web m
RDA includes some data for which there are controlled lists of terms or values - the Joint Steering Committee for Development of RDA considers almost all such lists to be open. In RDA we document the valid terms as they are established – or an individual bibliographic agency may choose to document their own terms for their user community – for example in a different language or establish an icon for the value- <click> like showing an image of a CD when the carrier type is a computer disc.
The terms used for the categories of entities established <click> in FRBR like work, expression, manifestation, and item – these terms are now registered through IFLA. The values or terms to be used for media types, content types and carrier types were developed in collaboration between the RDA editor and the publishing community that developed ONIX and are being registered on the web – most are also included in the DCMI/RDA Task Group’s registry of RDA elements with Dublin Core. <click> Similarly there are freely accessible registries on the web of controlled vocabularies, such as the Library of Congress Subject Headings (LCSH) in SKOS format (at <id.loc.gov/authorities> ) and the Virtual International Authority File (VIAF) for the names of persons, and we can expect more such registries. LC hopes to soon add LCC. 12 12

13. Internet “Cloud” Services Databases, Repositories Web front end
The registries are part of the cloud computing environment. In the cloud, data is available to be used by software that understands the connections and serves up context appropriate displays for users.
That is, by having the various names for things clustered together and identified, software can be written to extract the term or name that best suits a user’s specified preference for language or script and then can display that to the user. 13

14. VIAF Virtual International Authority File (creative re-use of data)
viaf.org
The Virtual International Authority File is an example of creative display of data coming from various sources that follow various standards.
Right now VIAF is personal name and corporate body name authority data plus uniform titles from about 25 institutions or consortia representing hundreds of libraries around the world.

15. The data from authority records is linked to data mined from the associated bibliographic records and packaged for VIAF to display, such as we see here with the places of publication for the works of the particular person, converted to a map of the world…or

16. The dates of publication from the associated bibliographic records, displayed as a timeline, a bar graph.

17. We also have other information in authority records, such as the primary language used by the person, the person’s nationality, and RDA explicitly labels these. In RDA such data is specifically identified and not just put into a general note. This makes the data more easily used by machines.

18. LCSH/SKOS and Visualizations
For another example of controlled vocabularies available on the Web, LC has put our LCSH vocabulary out on the Web and it has already been used by several systems , such as the European Library – TEL, to help provide vocabulary and references to get users to information they want. We are working towards connecting the various language versions of LCSH around the world, again to facilitate future switching of a user-preferred language and script for the LCSH terminology.

19. Database/format Scenarios
Based on Gordon Dunsire’s slide
Bib record (flat-file) Z
666.7
.L55
2009
Lee, T. B.
Cataloguing has a future
1 sound disc
Spoken word.
Donated by the author.
By focusing on specific identifying elements, RDA uses the FRBR structure to describe the resources in our collections. It enables us to prepare our bibliographic data for use in the Web environment.
Let me show you a brief demonstration of the re-use of the metadata and in particular the migration from card catalogs to MARC records and systems with linked bibliographic and authority files and then into FRBR structures and future linked data systems.
This is an animated view based on a slide from Gordon Dunsire formerly at the University of Strathclyde (Scotland).
In the past we had bibliographic data on a catalog card - when we transcribed data from the item, we typed or wrote it on the catalog card. The recording of the metadata was then used for displaying that data to our users in the card catalogs. We included such metadata as the author, title, content type, carrier type, subject terms, even possibly the provenance data as we see here (and of course a lot more).
1. Metadata 19 19 19 19

20. Database/format Scenarios
Based on Gordon Dunsire’s slide
Bib record (flat-file)
$a Lee, T. B.
$a Lectures on metadata
$a 1 sound disc
$a Spoken word.
$a Donated by the author.
$a Metadata
When we moved that data into the MARC format, we started creating machine-readable bibliographic records and authority records. But these were still basically meant to be stand-alone records that were self-contained and didn’t need referencing or links for data to be understood or indeed used to print catalog cards or display in local OPACs. 20 20 20 20

21. Database/format Scenarios
Based on Gordon Dunsire’s slide
FRBR registry (IFLA)
Future record
FRBR record
RDA element registry
Bib record (description)
Bib record (flat-file)
Work information
Name authority record
Author:
Lee, T. B.
Title:
Lectures on metadata
Name:
Work title:
Lectures on metadata
Content type:
Spoken word
Identifier: …
With some online systems even today, the authority data is actually stored in an <click> authority record and subject terms <click> are in subject authority records <click> with links between the bibliographic and authority records. We record the data in MARC format but it is displayed in a different way through our online catalogs to our users. <click> In a FRBR-based system, we separately identify <click> item level data, such as provenance of that particular item; <click> manifestation level data such as title proper and carrier type and the publication information such as the imprint – place of publication, publisher’s name, and date; <click> expression level data, such as content type; and <click> work level data, such as the <click> subject headings and the name of the creator of the work – and the work’s title, which in many cases is based on the title proper of the 1st manifestation. All of this data is linked and is used to identify each resource. It may be contained in a single package or through linked packages of data, depending on the system design. Future systems should make this structure invisible to users and easy for catalogers to create and maintain. <click> We also have the machine-actionable registries for the controlled vocabularies that we are using for <click> content type and <click> carrier type. In the future we may wish to link all names and terms to authority records or registries – even for names now in notes <click> or publisher’s statements. The terms we use for the roles, <click> like “author”, “composer”, “artist” – are also controlled vocabularies that we are making available as Web accessible registries. In fact, all of the RDA elements and sub-element terms are being put into registries on the Web. Likewise the <click> ONIX terms for content types are in a registry on the Web and <click>IFLA now has a namespace where you will be able to find all of the FRBR elements.
Expression information
Carrier type:
Audio disc
Subject authority record
Subject:
Metadata
Manifestation information
Provenance:
Donated by the author
Label:
Identifier: …
ONIX
RDA content type registry
Item information
Label:
Spoken word
Identifier: …
RDA carrier type registry 21 21 21 21

22. Linked Data Work information Name authority record Author: Subject:
Lee, T. B.
Work Title: Lectures on metadata
Identifier: …
Expression information
Subject authority record
Content type:
Manifestation information
Label:
Metadata
So my vision of the future is that this linked data would be available for re-use on the Web using the registries and repositories of description sets – done once and shared and maintained by all.
The structure would not look like this – it would probably be RDF triples, some with the URIs or universal resource identifiers for the specific values - and could be displayed using style sheets to fit whatever context was needed.
Catalogers won’t be asked to create 4 separate records for work, expression, manifestation, and item – they would be building the resource description as linked description sets, clearly labeling what belongs at which level and what is related to what, so machines can use the data, and can display to end-users in wonderful ways, depending on the task – whether displaying as a bibliography, or responding to a reference question for information, or whatever.
Title:
Lectures on metadata
Identifier: …
Carrier type:
RDA content type registry
Item information
Provenance:
Donated by the author
Label:
Spoken word
RDA carrier type registry
Identifier: … 22
Audio disc 22 22 22

23. Package for displays Future display ( ) Author: Lee, T. B.
( )
Author:
Lee, T. B.
Content type:
Spoken word
Title:
Lectures on metadata
Carrier type:
Audio disc
(click through)
We could even have the systems provide icons or other interesting devices to help our users quickly see the options of what’s available and display the data in a context-appropriate form.
The communication schema and coding with URIs or MARC tags would be invisible to both catalogers and end users – our library patrons. 23 23 23 23

24. Structures for Descriptions
Hamlet
México City
2008
English
Spanish
French
German
Shakespeare
Library of Congress
Copy 1
Green leather binding
Romeo and
Juliet
Stoppard
Rosencrantz & Guildenstern
Are Dead
Movies …
Shakespeare
Hamlet
Mexico City
We have become very accustomed over the past 40 years with the MARC format to think in terms of packaging the data about our resources and the associated entities as bib or authority records. Those bib and authority records have included identifying information as well as clues about significant bibliographic relationships. But that view of how we package the attributes will evolve as our information discovery tools mature.
For several years now, the National Library of Sweden has been experimenting and applying FRBR in systems with linked data <click> for a semantic web environment that seems to epitomize the direction we are heading. We want people to find things that are available to them, much like Amazon, Google, or any business trying to put their customers in touch with the products and services they have to offer. We have an inventory of resources that we need to describe and show how they inter-relate so our customers, our users, can find or learn about resources we have that will meet their information needs.
These things in our inventory – in our bibliographic universe – are described in the FRBR conceptual model as entities that have relationships and can be described by their attributes. RDA calls those attributes “elements”, to be in line with the semantic web and RDF (Resource Description Framework) structures and schemas.

25. Transformation of the Bibliographic Framework
Discussion list
Website
According to a recent press release: “The Library of Congress is launching a review of the bibliographic framework to better accommodate future needs. A major focus will be to determine a transition path for the MARC 21 exchange format in order to reap the benefits of newer technology while preserving a robust data exchange that has supported resource sharing and cataloging cost savings in recent decades. This work will be carried out in consultation…” worldwide with other libraries, agencies that provide library services and products, advisory committees and professional organizations, standards groups, and other experts.
There is an official Website as shown here and you can sign up to follow and participate in a discussion list at the address shown here.

27. Wish List
RDA Toolkit workflows for step-by-step approach for catalogers - with links to templates in local ILSs or bibliographic utilities
ILS links from specific elements in input screens to RDA instructions
The Joint Steering Committee for Development of RDA (JSC) has put a lot of effort into building the infrastructure and we still have a lot of work ahead of us, but what is still needed to reap the optimal benefits from RDA? So this begins my wish list, but it also includes some wishes from the JSC and others…
The RDA Toolkit includes workflows and a wizard to let people devise their own workflows which can be step by step procedures with links to the RDA instructions – or your own local documentation. Perhaps someday the workflows can also link to templates in integrated library systems or bibliographic utilities to help catalogers needing guidance for providing a specific bit of identifying information. <click>
I also hope that the integrated library system (ILS) Vendors and bibliographic utilities like OCLC will build in links to RDA so their input screens would link directly from the specific data elements to the RDA instructions – so we could go both directions when we are building bibliographic descriptions or authority data. OCLC’s Connexion now has such links to the RDA Toolkit. 27 27

28. Wish List Import descriptive metadata
Publisher/author supplied (e.g., ONIX)
Third-party supplied
Book vendors, contractors
Validation of required “core” elements linked to mode of issuance
I want systems that will easily import the descriptive metadata that comes from publishers (as OCLC and LC are now doing this with ONIX data) or for any metadata that accompanies digital objects - either literally as part of the digital object or data that is intended to go with the object but available separately, like thought a vendor service. For some of our digitized objects at the Library of Congress, we package both the digital object and its descriptive data – in a METS package. We should be able to have that descriptive data available for searching and displays.
<click>
I want to see our systems provide validation of the RDA core elements appropriate to the mode of issuance of the resource being described – to make sure the cataloger didn’t forget something. 28 28

29. Wish List Import controlled metadata
Registries for RDA/ONIX terms
VIAF (language/script appropriate to user)
Drop down menus for controlled vocabularies
Media, content, carrier types
Names of persons, families, corporate bodies
Subject headings
I want APIs (Application programming interfaces), that is, computer routines that provide the means to easily import controlled metadata from registries on the Web or international authority data systems, like VIAF – the Virtual International Authority File – that can also enable the display of languages and scripts appropriate to the user.
<click> I also want systems to provide drop down menus for the controlled vocabularies, such as the particular terms for specific RDA elements that have controlled lists of terms , for example, the forms of notation, content/media/and carrier types (like cartographic image, computer program, performed music, still image, text, three-dimensional moving image, etc, and computer discs, microfiche, volumes, videocassettes, etc.), and to suggest the appropriate links to names of related persons, families, or corporate bodies identified in the attributes and relationships; as well as suggest subject headings. In fact… 29 29

30. Wish List
Automatic suggestion of classification/subject headings for works
Based on keywords found in resource record and/or digital resource itself or accompanying tables of contents, abstracts, summaries, etc.
Based on matches with existing similar works
Classification and subject heading correlations
I want to see the automatic suggestion of subject headings and classification numbers for LCC, Dewey, and other systems that wish to register their schemes <click> based on what is included in the description sets and linked digital tables of contents, summaries, abstracts or text. Those suggestions then could be confirmed or corrected or added to by a cataloger. The National Library of Australia has a “subject suggester” system that does this and the Library of Congress is experimenting with a similar system for suggesting Library of Congress Subject Headings.
I’d like to see suggestions <click> made based on matches with existing similar works as well as with correlations <click> between classification and subject heading systems, as we now can do in Classification Web. 30 30

31. Wish List
Automatic generation of work/expression data and links to “creator”
Based on identifying elements for the “first” manifestation
Automatic prompting and validation of work/expression data
Suggests possible matches for new cataloging
Also on my wish list….Rather than catalogers manually keying in separate authority records for all the works and expressions…, have them automatically generated, because the description set of relevant metadata <click> should mostly already be there or be generated based on the first manifestation received, and systems should suggest links that catalogers would verify or correct – saving a lot of time and effort. We know there are many expressions of works for various editions and translations or performances for music, so when we build the description and access points for a resource, the elements that identify the manifestation with its associated work and expression can be used by our future computer systems to automatically build the descriptive data for the name of the work and the relationship to the creator of the work, as well as provide a placeholder date of the work/expression. When we build a MARC record now, we add the language, which is expression data, and future systems should know to label it as expression data.
Our future systems should automatically prompt us of existing <click>work/expression data. The data from the new manifestation should be matched, and exact or fuzzy matches should be displayed to prompt the cataloger to verify it’s the same work or expression. The computer systems should <click> suggest relationships and make it easy to designate a relationship, such as with a touch screen, where you point to make the link; and default to general types of relationships with the option to select more specific ones as needed or add new ones to a controlled list of designation type terms. 31 31

32. Wish List
Easy way to share “maintained” data worldwide (from “On the Record”)
And I want to see catalogers have an easy way to share data that is maintained cooperatively – a cataloger should not have to feel they must provide everything – other people with the information at hand or with the expertise should be able to enhance the shared data. This was a goal that was also mentioned in “On the Record” by the Working Group on the Future of Bibliographic Control”. When the data package or description set is shared, it’s there for everyone to use in displays.
Or there can be automatic updates from the shared data to populate local systems that choose to keep their own data redundantly. 32 32

33. Wish List FRBR collocating (expand and collapse elements for displays)
Simple displays of pathways to related resources and information about related entities
Open options if user wishes to explore – don’t overwhelm with all possible relationships
For our end users including catalogers, I want to see more FRBR-based systems that make it simple and easy to expand or collapse the set of elements that are displayed for collocated entities and show the relationships to help them easily move about a surrogate bibliographic universe – but to do it in a way that does not overwhelm them with graphics or text - a way that let’s them be in control of how much or how little or which direction they take to explore the available resources.
And I’m sure you can think of lots more for your own wish list! 33 33

34. Where are we headed?
We have seen the expansion of global search engines and experiments with the semantic web that hold great promise for helping us move from the idea of needing a catalog for our library to instead seeing how the collections in our libraries can be more accessible by users anywhere at anytime by using our catalog records as a trusted description of the resources we have to offer.
Given the library investment in our integrated library systems and bibliographic utilities, we probably cannot make the jump directly to free global shared description and access, but RDA is starting us on a path to go in that direction. And for several years to come we will be in a transition period, or what I call a “bridge” period, because we are not yet able to totally break with the past and move on.
We need a really big change to design totally new structures and systems. We need to move on to a day when there are no headings, but just the various names for things that can be displayed in whatever language or script a user wants, and there are no bibliographic records, but just description sets of resources that are linked to other related resources…that will take some time unless we see another Google-type movement that shows the world how to do it. So where are we headed? (See slide)
Let’s assume for now that the semantic web of linked data in a cloud computing environment is a good next step. We start with providing our resource descriptions in a well-formed structure that can be easily used in such a linked environment.

35. Resource Discovery System
User-focus
Builds on existing descriptive metadata clearly labeled
Identifies all names and other identifying information for an entity
Identifies significant relationships to enable collocation and navigation of the bibliographic universe
Re-uses data globally for more efficient operations
That vision of the future involves a user- focused resource discovery system tied to local inventory systems that uses existing descriptive metadata from the resource itself as much as possible with lots of relationships. We should have systems that cluster together the identifying data about each entity, such as variant names to increase the recall of relevant matches and to increase the precision of searches by clustering the distinguishing information about an entity. We should have systems that re-use the identifying data about entities and their relationships for more efficient operations, more creative displays of information to answer many information questions, and to offer pathways to related resources in our bibliographic universe.
We get there by providing a clearly labeled set of data elements for the entities and relationships that matter to us in the bibliographic universe – the things in our collections and the associated people, corporate bodies, and families, and the various relationships, including subject relationships that help us connect our users to information we have for them – whether they are looking for something specific (a known item search) or need help. We can inform them about related resources or information that they may find relevant to their information needs.
And to do this in a way that builds on data that is readily available.
RDA is helping us move in that direction. 35

36. Considerations to get there
Bridge
What incentives are needed?
How would staff of different institutions, programs, publishers, systems, national, etc., function in and support that scenario?
What are the economic/legal obstacles to overcome if some controlled vocabularies are proprietary, restricted, or less easily available on the Web?
Does someone need to start the ball rolling? Who? How?
And how do we get to that visionary scenario? I did mention that we have some things for the bridge period where we are still using MARC records.
But here are some of the further questions that we’ll need to think through together: <click>
what are the incentives for institutions to reach that future? There will be cost savings for the cataloging operation and service providers through shared linked data, once we can get systems that are built on this vision, especially when there is no need to ‘exchange’ records or download all the bibliographic and authority data locally but we can share it globally. However, we also need to leave open the option for those institutions that want to continue our current models of redundant data at each locality.
We need to share this story with library directors to help them better understand and build a common purpose. <click>
We need to imagine how staff from different institutions including libraries, archives, publishers, and distributors would function to support this vision, what would be the role of national bibliographic agencies beyond maintaining controlled vocabularies? <click> How many of those institutions would be able to make their vocabularies freely available and what mechanisms need to be in place for any models that would require payment for data. <click> Who should start the ball rolling and how? 36

37. FRBR-Based Collocation Linked Data
Stoppard
Shakespeare
Derivative
works
Hamlet
I look forward to discussions in the near future that will help us move closer to better systems to help users – to collocate related resources and to offer the pathways for users to explore our collections and the resources available to them worldwide so they can find what they need in global web-based systems.
We may need to begin with small steps to make the best use of existing MARC record structures, but hopefully that will position us for future structures and systems that take full advantage of the model for streamlined cataloging operations and more effective user service.
The first release of RDA reflects many compromises that carry over some of the traditions of case law from AACR2, but that is intentional so there will not be too much change, because we are being told by administrators to avoid the trauma that many of us still remember when we moved to AACR2, from the refusal to change old practices of corporate entry under place names – better known as “superimposition” for AACR, and then the greater trauma of “de-superimposition” when we moved to AACR2 that caused the splitting and closing of card catalogs. There was something to be said for a huge step like de-superimposition because it spawned the development of online catalogs, but unfortunately they were only extensions of card catalogs on a computer, and still have not realized the hopes held for them. We have limped along with online catalogs for too long and have seen Google and Yahoo and Amazon at work in a cloud computing environment with linked data and a global perspective, building a rich and user friendly information community.
With RDA, the JSC has engaged various communities and worked together towards a more principle-based standard for describing our resources – a standard that results in well-formed metadata that is re-usable globally. A lot of work still needs to be done, but it’s a step towards better service to our users.
Thank you for your attention. (Note next slides)
Rosencrantz & Guildenstern
Are Dead
Romeo and
Juliet
English
French
Text
Movies …
German
Swedish
Subject
Stockholm
2008
Library of Congress
Copy 1
Green leather binding 37 37 37 37 37

38. Links http://loc.gov/aba/rda/ LChelp4rda@loc.gov
LC information on RDA Implementation Planning
LC Webcasts, training materials, examples, test records, etc.
and 38

39. Links RDA Database Implementation Scenarios
Encoding RDA data
RDA, FRBR/FRAD, and Implementation Scenarios
MARC development web site
DCMI/RDA Task Group wiki
XC Project – recent description by Jennifer Bowen 39 39

40. More Links IFLA FRBR http://www.ifla.org/VII/s13/frbr/ FRAD
ICP
JSC
Also see the RDA Bibliography posted on the Library of Congress site: 40