Panel Participants

Unique Identifiers: DOIs, PURLs, URLs, URNs, RFCs, RDFs, R2D2, ? or ? Are You Asleep Yet?
Elizabeth B. Richmond, University of Wisconsin - Eau Claire

LPS Cataloging Policies for Internet Resources: A Review.
Thomas A. Downing, Chief, Cataloging Branch, GPO

Building the FDLP Electronic Collection: A Policy and Planning Document.
Laurie Beyer Hall, Library Systems Analyst, GPO

Respondents
Nan Myers, Wichita State University
Arlene Weible, Willamette University
Tim Byrne, University of Colorado, Boulder

Discussion following the presentations

Presentations

Unique Identifiers: DOIs, PURLs, URLs, URNs, RFCs, RDFs, R2D2, ? or ? Are You Asleep Yet?
Elizabeth Richmond

Following is the text of her PowerPoint presentation: bibliography is also available.

Objectives - to provide an overview of:

• Unique Identifiers for document-like objects/digital content objects
• User needs
• Uses and requirements of a standard identifier
• Unique Identifier solutions
• Conclusions?

Content identification

• The Digital Object Identifier (DOI) system was designed to overcome the impermanence of Uniform Resource Locators (URLs) to identify and track digital information at any level of granularity (detail) such as an article, a data set, a book, a table.
• The location and retrieval issues with URLs, and to some extent with PURLs (Persistent URLs) is that they link to a site, a network location, not to a specific content.  This is analogous to locating a resource’s placement on a shelf or at a particular branch library.  If the resource is moved and only the previous location is identified by a retrieval system, then users’ needs aren’t met.

•  Overload
• More is less
• Knowledge is power (Francis Bacon)
• Where is the information leading to the knowledge?
• How can I identify the content I want (manifestations and versions)?
• Need for unique identifier standards
• How can I identify the content I want (manifestations and versions)?
• Manifestations: different expressions or formats of content (SGMl, .pdf, .txt)
• Versions: content revision
• Limitations of print legacy identifiers (ISBN, ISSN, SICI ?)
• electronic information has more convolutions than the pre-printed page.

• Primary commerce (such as uses of an article or database)
• Rights tracking (similar to commerce)
• Identification and location
• Cataloging
• Citation

1. Uniqueness - identify content uniquely and persistently

• e.g. STM journal article may need identifier for article and abstract
• annual editions must have unique identifiers

• URLs make reference to materials on the internet
• URNs (Uniform Resource Names) more effectively serve as object identifiers
• DOIs are independent of transport protocols such as TCP/IP

• Different types of content and media
• Different levels of granularity (as metadata)
• Different manifestation & content versions
• Change of ownership

Unique Identifier “solutions”

• URLs: identify network location, not content
• PURLs and URNs Names: Internet/WWW partial solution to identifying content rather than simply pointing to a network location.
• Other legacy identifiers:
• ISBN and ISSN
• BICI (Book Item and Contribution Identification) and SICI (Serial Item and Contribution Identifier) [NINO standards]

• Competing or complementary technologies?
• Will PURLs support unique identification of digital content objects?
• What will IETF support?
• URNs?
• DOIs?
• BICI?
• SICI?

LPS Cataloging Policies for Internet Resources: A Review.
Thomas A. Downing

The main points of his talk follow.  A complete version was published in Administrative Notes,  available at http://www.access.gpo.gov/su_docs/fdlp/pubs/adnotes/ad022599.html#4.

Beginning about four years ago, the initial effort was to record URLs in the 530 note fields for serials.  Some of these initial efforts were not well executed, with URLs leading to such inappropriate sites as advertisements for publications.  As time passed, URLs led to specific titles and issues, and ways to find them (through indexes).  In 1996, a policy was proposed to the Depository Library Council (DLC) for the creation of very brief records which would consist of a title, series (if available), SuDocs class number, and hot-linked URLs.  This was intended to provide low-cost access, especially in view of the all too common practice of many agencies of taking the titles down after a brief time.  DLC favored full AACR2 cataloging so that titles could be included in online catalogs, rather than some kind of stand alone access.  Efforts to provide additional approaches to Internet resources have resulted in a suite of pathway services, including Global Information Locator Services (GILS).

At approximately the same time, CONSER members and OCLC personnel were developing policies and practices which provided a national-level context for Internet cataloging initiatives.  For their part, CONSER members advocated the use of a single record approach, as an option to creating separate records for the various formats.  As members of CONSER, GPO was authorized to use the single record option for cataloging of Internet resource parallel to the microfiche or paper records.  LPS records provide access to Internet resources through the use of physical format records for other formats.  Physical format records may be adapted for use with Internet resources, and Internet resource records may be adapted to describe physical formats.  Use of a single record to record different formats represented a major savings.  CONSER also advocated the use of the 856 field for the URL, allowing hot linking from records to catalogs, saving time and effort.

The application of CONSER related policies does not solve the problem.  The OCLC PURL software which provides a seamless redirect function, does ease some of the maintenance for GPO and libraries.  There are now over 2200 PURLs being maintained by GPO, and the software has improved.  PURLs are a means to an end.  GPO currently has 2 full time employees maintaining PURLs, and it may take more.  Personnel are engaged in the following important tasks: choosing the best URL; reviewing exception reports for broken links; making sure that the new URL is an official version.  Maintaining PURLs requires human intervention and professional judgement.  GPO’s PURL applications support the Browse Electronic Titles (BET) and Monthly Catalog of United States Government Publications (MOCAT), Internet cataloging and indexing.  BET entries essentially represent GPO’s initial proposal for short records.

Physical format records are upgraded to provide access to electronic resources with PURLs or URLs put in 856 fields, and a note added to indicate the electronic access.  When no suitable record is available, an “electronic only” record may be created to provide access.  These may be later upgraded to indicate availability of the information in a physical format.  It can work both ways.

In 1998, GPO has also implemented use of a “collected works” record for collections or groups of files accessible from a single web site.  This was endorsed by the GODORT Cataloging Committee with assurances that GPO will not use this unnecessarily, but when the composition of the files available from the site change frequently, these records will help even while single titles are also being cataloged.  GPO works with the Library of Congress and others to provide as much access to electronic works as possible.  This policy is subject to review.  Application of this policy allows LPS the flexibility to provide aggregate cataloging records for Internet sites and products that are valuable to users but which cannot be cost effectively cataloged at the piece level.

To assist with efforts to enter a new phase of services in the context of an FDLP electronic collection, Laurie Hall and Julie Andrews are working with depository libraries and other GPO branches to work on the collection and electronic archives.  New thinking and continued policies will be developed in the move to an increasingly electronic FDLP.  The two new concepts are an electronic collection and archiving.

Building the FDLP Electronic Collection: A Policy and Planning Document.
Laurie B. Hall

Following are points from her PowerPoint presentation. The text is also available in Administrative Notes v.20, no.5 (February 5, 1999) available at http://www.access.gpo.gov/su_docs/fdlp/pubs/adnotes/ad022599.html#5.

Workgroup Goals for Building the FDLP Electronic Collection

• Provide permanent public access to Government information
• Recognize the reference needs of our user community
• Provide quality bibliographic control of Government information

• Managing the FDLP Electronic Collection
• Title 44, Chapters 19 and 41, U.S. Code

1. Core legislative and regulatory GPO Access products which will reside permanently on GPO servers
2. Other remotely accessible products managed by either GPO or other institutions with which GPO has established formal agreements
3. Remotely accessible electronic Government information products that GPO identifies, describes and links to but which remain under the control of the originating agencies
4. Tangible electronic Government information products distributed to Federal depository libraries

• Enlarge and Improve Communications
• Refine Criteria for Selection
• Refine Criteria for Analyzing Resources Selected for the Collection
• Test Processing Decisions
• Investigate Storage Options
• Evaluate Locator Services
• Investigate Collection Management implications

• FDLP Community
• Agencies
• Peer Institutions (NARA, NLM, EADS, etc.)
• GPO Offices

• Title 44, Chapter 19, U.S. Code
• Official, authentic Government Information
• Presents a major activity of the agency, or is a product of a major activity of the agency
• Not selected:
• Official Use Only
• Administrative Use

• Is this resource an important product for the FDLP user community?
• Is this suitable for the FDLP user community?
• What level of bibliographic control would best fit this resource?
• What locator service will be best suited for this resource?

• Is this in scope for the FDLP?
• What type of product is this?
• Should this be in GILS?
• Should this be in Browse Topics?
• Should this be on BET (Browse Electronic Titles)
• Should this have a PURL?
• Should this be archived?
• How should this be cataloged?
• New or adapted?
• Level of cataloging?

• Goal: Provide a limited number of tools that are easy to use, easy to maintain and give the best possible access to the electronic resources.
• Address issues presented by focus groups and community feedback
• Examine ways to merge and consolidate
• Test format and presentation changes
• Solicit partners

• GPO Access
• GPO Archive
• Agency Resources
• Partners

• Hire a Collection Manager
• Future Decisions for Collection Manager and Staff:
• Personnel
• Maintenance
• Resource Allocation
• Promote Cooperation with Partners

• Communications with all concerned stakeholders
• Consistent processing procedures
• Permanent public access to the FDLP Electronic Collection
• Locator services that are easy to use and manage
• An integrated library system

Discussion following the presentations

Question: GPO’s use of the single record option creates Interlibrary Loan request problems for libraries whose holdings symbol is attached to the record in the OCLC database.  These libraries don’t have the back issues to loan, merely access to the electronic versions on the web.
Downing: When it comes to the single record approach, institutions do have very legitimate needs and differing concerns about this.  What we’ve tried to do, and what we generally try to do, is to follow what is authorized in terms of national-level policies, realizing, unfortunately, that for some libraries this may create a problem.  We have tried to take the normal, anecdotal surveys, and what I have found is that a surprising number of technical services librarians favor a single record approach.  I can’t make a statement that this is absolutely true — that there are too many variables — but what this means for us, in terms of resources, is that we can handle the work that comes into us... with very limited resources.  If a consensus emerges, that this is not acceptable, then we will have to take a look at things.
Additional Comments:  Several people expressed strong support for the single record approach, from both public service and technical services perspectives.

Question: A problem with PURLs is that link checking software frequently won’t recognize the bad links because of the PURLs “self resolve” feature.  Even if it never takes the user to the site, link checkers see it as a valid page.
Downing: OCLC is looking into this problem because there is a need for link checking.  LPS is open to whatever technology is recommended.
Weible: One of the features of PURL software is link checking.  We will have to trust that GPO is checking PURLs, but many libraries want to check their own links.  What is the most effective way to report broken links?
Downing: askLPS.

Question: How often does GPO check links?
Downing: Weekly.

Question: How can libraries share the lists they have made of the broken links?
Downing: Send the lists to askLPS for GPO to fix.

Question: What is the time period between the reporting of a URL problem or correction to askLPS and GPO’s passing along the information?
Downing: The askLPS service is caught up now.

Question: Some records show a title cataloged as a monograph, but new editions are found when the site is visited which do not match the older edition described.
Downing: GPO is doing all they can with its limited resources, but is not evaluating past work.  If something disappears, GPO catalogers are adding a “no longer available via the Internet” note.
Richmond: Part of the problem is that these URLs are based on the network location rather than on the content.  The solutions are based on the wrong questions.

Question: Some links are broken because of the actions of producing agencies.  How are conversations with publishing agencies going?
Downing: The back files need to be put in one place or archive.  GPO is archiving the titles on GPO Access and is working with agencies to assure that files are permanently  protected until GPO can archive them.
Hall:  One of the most effective ways of communicating to agencies is for the users to contact the agency directly with their concerns [about an edition being removed from a site, etc.] Hall then stated that comments to agencies from users are more effective than GPO contacts, in her opinion.

Question: The use of multiple subfield u (in an 856 MARC field) is a current practice.  Will this be continued?
Answer: Consensus is that only a single subfield u is allowable per 856 field.  This represents a change from the use of a multiple subfield u’s.

Question: When GPO announces new electronic resources (coded as (EL) in Administrative Notes Technical Supplement) could their PURLs be published instead of URLs?
Hall:  This can be done if the PURL is known at the time of the announcement.
Byrne:  Updated records require updates in many catalogs.  Subscribers to record services have heavy maintenance issues.  Some type of notification to libraries about the assignment of PURLs, or changes to URLs, is very important.  Libraries are facing heavy maintenance issues, whether they subscribe to record services for tapeloads or update records in other ways.
Downing: What do people want to allow maximum utilization of limited resources?"
Myers:  A method to assist batch loaders and updaters.  A PURL listing alone would not be helpful without the context of some kind of identification such as the OCLC accession number.
Downing: GPO would welcome feedback from the GODORT Cataloging Committee about PURLs linked to a unique identifier.

Question: Is it possible to purchase records for electronic titles only?
Answer: (From vendor) Yes.
Myers:  Work is in progress on a DataMiner option to select the data from 856 fields.

Weible: In regard to the section of Hall's presentation on "Processing Decisions", how will GPO determine the appropriate access point (GILS, BET, Browse Topics, etc.) for each resource in its Electronic Collection?  What criteria will be used? Will there be a minimum level of access for all resources? For example, will all resources have at least a catalog record, with other access points as appropriate?
Hall:  Maybe not, as some could be too general.  Not all BET titles will have cataloging records.
Weible: I would urge GPO to allow FDLP community to be involved in the development of criteria for establishing appropriate access points, as this is where depository librarians have a lot to contribute.
 

Return to Cataloging Committee Home Page

Updated 3/22/99