The U.S. Government Publishing Office (GPO) continues to monitor the discussion of the Joint Steering Committee for the Development of RDA: Resource Description and Access. RDA is based on FRBR, or Functional Requirements for Bibliographic Records, and was released in 2010. It has undergone testing at the Library of Congress, the National Library of Medicine, the National Agricultural Library, and at RDA National Test Partner Libraries.

In June of 2011 several updates were released, which add further to this discussion. The links to the full text of the documentation are included here:

• Report and Recommendations of the U.S. RDA Test Coordinating Committee
• Executive Summary of the U.S. RDA Test Coordinating Committee
• Response of the Library of Congress, the National Agricultural Library, and the National Library of Medicine to the U.S. RDA Test Coordinating Committee

The transition to RDA is complex, but several changes are notable in how they differ from current AACR2 standards:

• RDA will impose new rules for the creation of added entries (elimination of the "rule of 3").
• Most abbreviations are to be eliminated, particularly Latin abbreviations.
• The general material designation (GMD) will be replaced by three new MARC fields: content type (336), media type (337), and carrier type (338).
• RDA will use mode of issuance instead of continuing resources.
• The impact of RDA on authorities. Much is still unknown about the impact of RDA and the hybrid environment in the national authority file, although a policy has been developed for PCC members on using RDA and AACR2 authority records.

The general recommendation of U.S. RDA Test Coordinating Committee regarding the timeline for completion has been revised, with a projected implementation date no earlier than January 2013.

Progress Updates

Update 10 (9/15/2015)

GPO successfully completed implementation of RDA in April 2013 and now catalogs new bibliographic and authority records according to the RDA standard. In FY15, GPO continued to monitor and apply new and updated practices in RDA, its Library of Congress Program for Cooperative Cataloging Policy Statements, and other documentation of the Program for Cooperative Cataloging (PCC). In addition, Library Technical Services continued to review, evaluate, update, and develop local practices and procedures. As a result, nine new GPO Cataloging Guidelines were published in September 2015. The full complement of remaining bibliographic GPO Cataloging Guidelines will be published in FY16. Finally, in FY15, Library Technical Services trained new staff in RDA practices, offered refresher training to existing staff, provided personal training and improved staff performance through ongoing quality assurance, and presented eight technical services webcasts and webinars to the public.

Future updates to GPO’s cataloging practices will be posted via the FDLP News & Events email services and on the FDLP home page.

Update 9 (2/12/2014)

The RDA Implementation Team’s quality control of Congressional titles revealed three issues that were addressed with modifications to GPO cataloging practice. Beginning in December 2013, GPO Bibliographic Control and Metadata librarians use the following practices for newly cataloged Congressional titles:
 

1. The application of relationship designators germane to these document types, including enacting jurisdiction and jurisdiction governed.
   
   Decision: GPO will apply enacting jurisdiction to the United States for public and private laws, statutes, and treaties.
   
   Decision: Jurisdiction governed will not be applied immediately. Congressional catalogers require that this term be investigated for determination of its scope and precedents before application
    
2. Differentiation of titles for different works with identical authorized access points (AAPs), per RDA 6.27.1.9 and its LC-PCC PS (monographs section)
   
   Decision: For recurrent Congressional hearings with identical AAPs, GPO will apply the full date of the hearing in the parenthetical qualifier.
   
   Decision: For recurrent Congressional reports with identical AAPs, GPO will apply the publication year in the parenthetical qualifier (until/unless this decision fails to effect adequate differentiation over time).
   
   [Treaties already are covered by: RDA 6.20.3.3: Recording Date of Signing of a Treaty, Etc., and RDA 6.29.1.33: Additions to Access Points Representing Treaties, Etc.]
    
3. Library of Congress Genre/Form Terms (LCGFTs) will be applied to Congressional publications when possible. From the LCGFT list, the broad term Law materials is an excellent starting point for discovery of applicable terms. Frequently applied terms in MARC records thus far include:

•  
  • 655 _7 Legislative materials. $2 lcgft
  • 655 _7 Legislative hearings. $2 lcgft
  • 655 _7 Statutes and codes. $2 lcgft

For more information about these practices, a webinar: Congressional Publications: An Overview, will be broadcast on Thursday, February 20, at 2:00 p.m. EST. 

Update 8 (12/23/2013)

The RDA Implementation Team continues to perform quality control (QC) checks of GPO’s bibliographic records for conformance with RDA instructions and for articulation of local policies. In September, all catalogers were briefed on the application of Library of Congress Genre/Form Terms (LCGFT). Bibliographic records for audio-visual resources and cartographic materials now contain LCGFT headings.

LCGFTs are coded in MARC as

• 655 _7 |a [term] |2 lcgft.
• E.g.,
• 655 _7 |a Internet videos. |2 lcgft

LCGFTs applied for audio-visual resources cataloged by GPO include:

• Internet videos
• Video recordings *
• Video recordings for the hearing impaired
• Documentary films
• Sponsored films
• Public service announcements (Motion pictures)
• Radio public service announcements
• Podcasts
• Sound recordings *
• Audiobooks
• Legislative materials
• * Apply to collections per instructions in LCGFT thesaurus.

LCGFTs applied for cartographic resources thus far indicate specific types of maps:

• Road maps : Use for Motor vehicle use maps
• Nautical charts : Use for nautical charts from NOAA
• Quadrangle maps : Use for USGS quadrangles.
• Tourist maps : Use for maps from the National Park Service, National Forest Service, Bureau of Land Management and other agencies that show trails, recreation sites, and other features of interest to visitors.

An example of a recently created bibliographic record with LCGFT can be found in the Catalog of U.S. Government Publications (http://catalog.gpo.gov) is highlighted below. Additional examples are: OCLC # 855212750 / CGP system number 000906168; OCLC # 847855339 / CGP 000894331; OCLC # 727069179 / CGP system number 000820231.

OCLC # 841960920 / CGP system number 000898643
 

Update 7 (8/28/2013)

The RDA Implementation Team began quality control exercises for all bibliographic records created and updated since LSCM staff began cataloging using the RDA standards in January 2013. The team is evaluating records for consistent practice in MARC record coding and application of RDA instructions and LC-PCC policies. The first quality control review identified some inconsistent coding errors in the MARC fixed fields. All errors identified during this first exercise were corrected as of August 23, with all records updated in both OCLC and the Catalog of U.S. Government Publications (CGP). Additionally, clarifications have been issued to LSCM staff regarding relationship designators for issuing bodies and publishers, and for treatment of microfiche in retrospective cataloging. These clarifications will now be incorporated in newly created and updated bibliographic records and will be documented in the next revision to the GPO Cataloging Guidelines, currently underway. The Team also continues to comprehensively review authority records created in RDA. Exercises and data clean-up, if required, are ongoing and progress will continue to be reported.

Update 6 (5/1/2013)

GPO has been declared independent to create bibliographic records in RDA by NACO and BIBCO. Therefore, effective 1 April, 2013, all of GPO’s original bibliographic and authority records are being created according to the RDA standard.

In accordance with RDA policies articulated for copy cataloging by PCC and OCLC, GPO will continue to adopt existing AACR2 records when they are encoded at PCC level. A live training Webinar, "An Overview of RDA for FDLP Libraries," was offered on Tuesday, April 16, 2013 at 2:00 pm EDT. A recording of this Webinar is available for those who could not attend.

Update 5 (1/24/2013)

The following list is an update of tasks accomplished and training scheduled that will lead to the complete adoption of RDA standards for cataloging.

• GPO catalogers have completed NACO and BIBCO RDA bridge training, and have been declared RDA independent by Library of Congress PCC reviewers. Authority and bibliographic records are now being created using RDA standards.
• CONSER RDA bridge training will take place in February 2013 via three Webinars.

Update 4 (12/18/2012)

The following is a timeline of upcoming RDA tasks leading up to the complete adoption of RDA standards for cataloging.

• December 2012: Test batch of approximately 30 records created.
• December 20, 2012: Test batch records and all records created in RDA will be available for searching in both CGP and OCLC. A quick link to the CGP records is available on the CGP homepage.
• January-March 2013: Authority and bibliographic records created using both RDA and AACR2.  All records created in RDA and available in the CGP and OCLC, must pass newly developed internal quality control standards. RDA record creation will focus on high priority cataloging output, ie Congressional publications, electronic/online documents, ‘hot topics’, etc. Order records and brief acquisitions records will continue to be available in AACR2.
• April 2013: Full implementation of RDA is expected. All authority and bibliographic (full cataloging and brief acquisitions records) created by GPO follow  RDA standards.
• April 2013: A virtual information session about RDA will be held for depository libraries shortly after implementation takes place.

Update 3 (7/17/2012)

In June 2011, OCLC issued their RDA policy, OCLC Policy Statement on RDA Cataloging in WorldCat for the U.S. Testing Period and Beyond, which states:

“For materials other than continuing resources, catalogers are asked NOT to edit a WorldCat full-level master record (ELvl values blank, 1, 4, 8, I, and L) to change it from one set of rules to another. In other words, if the record was created according to AACR2 (and coded as such), please do not change the master record to RDA. If the record was created according to RDA (and coded as such), please do not change the master record to AACR2.”  

To date, GPO catalogs following AACR2, and does not yet catalog according to RDA (other than test records which have been cited in these updates). Since GPO follows AACR2, when we find records cataloged according to RDA in OCLC, GPO staff are locally editing records to conform to AACR2. When we produce them for the record load and the CGP, the records are cataloged according to AACR2, although the master record remains cataloged according to RDA in OCLC. GPO catalogers will remove all RDA elements from the local record, including the $e rda from the 040.

sys000860080  ocm767644750 -- C 55.302:H 11 -- NOAA Habitat Program

LTIS staff continue to follow the progress of RDA implementation. NACO training for RDA will be made available later this fall. No Day One for bibliographic records has yet been announced; GPO continues to plan for full implementation of RDA in 2013. All comments, questions and suggestions about this policy or its application should be submitted to Fang Gao, Supervisory Librarian, Bibliographic Control and Metadata Section, Library Technical Services.

Update 2 (10/6/2011)

In early September, the CGP was updated to accommodate the display of RDA information. The new MARC fields 336 (content type), 337 (media type) and 338 (carrier type) display in the CGP MARC view in numerical order. The 338 tag containing format information now displays in the CGP Standard view with the label “Format”. Check out Aleph record number 000822999.

Additionally, GPO finalized a local practice option for capitalization of title information in the 245 MARC field.

LTIS staff continue to follow the progress of RDA implementation. New titles have been cataloged following RDA principals. You can view these records in the CGP; system numbers 000832680 and 000829099. GPO is planning for full implementation of RDA in spring 2013.

Update 1 (7/12/2011)

Supervisors, managers, and Bibliographic Control staff continue their training on these new RDA standards. Bibliographic Control staff have recently upgraded several RDA records as part of their routine cataloging activity. These records can be viewed in the Catalog of U.S. Government Publications (CGP) (catalog.gpo.gov). The Aleph system numbers for these records are 000818552 (OCLC # 728302331) and 000822999 (OCLC # 728234001). Look in the MARC view of the CGP records to see the new fields 336, 337, and 338 and to see “rda” in the 040 field. An analysis of the RDA features of these records will be posted to this project page in the coming weeks.

Print on demand (POD) is a printing technology that allows books to be produced one or two at a time or in small lots, rather than in large offset print runs.

The U.S. Government Printing Office (GPO) explored the use of POD technology to broaden its title offerings, to ensure lasting availability of publications, and to avoid the heavy expense associated with surplus inventory.

Benefits to Federal Depository Libraries

For over a year, GPO discussed the possibility of providing a POD allowance for depository libraries to be used as a tool to augment depository library collections with paper versions of depository titles. GPO is still in the process of evaluating different options but one possible scenario is that the depository library could use an allowance to quickly order and receive a printed copy of a publication regardless of its item number. The policies being developed for use of a POD allowance may permit libraries to order multiple copies, replacement copies, or to supplement item selections with print publications that are of interest. A POD allowance might make the selection process of some print titles from GPO more like using a book dealer. For example, libraries would not have to guess in advance what titles will be issued under certain item numbers and then take everything issued under those item numbers. They could wait until the items are issued and then order only the titles they actually want.

POD technology was also examined as a way to produce tangible copies of titles that are not otherwise available to the Federal depository library community in print. These items would have to be suitable for replication using POD technology, which means that they have to be within certain standard sizes, not contain excessive color or artwork (such as a coffee table art book), or any other attributes that are not suitable for the POD equipment. Since many of these titles would not otherwise be available to Federal Depository Libraries except in electronic form, POD would provide an option for obtaining printed copies for a tangible collection.

GPO tested potential vendors to see exactly what capabilities are practicable now, as well as those coming in the future. The results were encouraging in most cases, and the ability to deliver quality reproductions through POD is available to take advantage of right now. The goal for the application of POD in the FDLP would be to offer POD publications as close to the original quality as possible, without sacrificing the ability to use the information. Quality access to authentic Government information remains an essential goal to GPO and will always be considered in the deliberations of how POD could be used to benefit the FDLP.

Survey Results

GPO conducted a survey of Federal depository libraries from September 8-27, 2004 to help gauge the future demand for print on demand (POD) publications. The survey presented two different hypothetical scenarios and enabled GPO to gather data regarding the level of interest for a representative sample of potential POD publications. For the purposes of the survey, the cost of each title was the estimated POD sale price, less a 25% (book dealer discount). To complete the survey, GPO asked depository librarians how many copies of each publication they would choose for their depositories under the following two hypothetical scenarios:

1. How many copies of each publication would the depository select if each depository was given an annual allowance of $500.00 for POD publications?
2. How many copies of each publication would the depository purchase with its own funds after the annual allowance of $500.00 was exhausted?

Survey Results

The Print on Demand Survey received a total of 402 unique responses from depository libraries which represents roughly 31% of the total number of libraries in the depository community. A representative sampling of 100 titles was listed on the survey in order to determine the types of items that libraries might have an interest in if they were printed on demand. Every title on the survey generated some selection activity from library respondents. For some of the titles, libraries expressed an interest in acquiring multiple copies for their collections.

Since March 1998, the U.S. Government Printing Office (GPO) has used persistent uniform resource locators (PURLs) to provide libraries and other parties stable URLs to online Federal information. GPO's PURL implementation is utilized by Federal depository libraries, Federal agencies, and others to provide persistent links to Federal resources.

In August 2009, GPO's PURL server experienced a significant hardware failure and was unavailable for an extended period. When this happened, all PURL links became inaccessible. The PURL system is a critical tool of the Federal Depository Library Program that fulfills its mission of providing permanent public access to Federal information resources. As a result of the outage, GPO took steps to ensure high availability and redundancy of the PURL application.

Rearchitecture of PURL Software and Hardware

In order to ensure high availability and redundancy, GPO migrated from the original OCLC PURL Resolver Software to the PURLZ Resolver Software, which is a re-architecture of the original OCLC PURL Resolver Software. Included with the migration is the implementation of an improved hosting solution that provides higher availability for the PURL application. Previously, (ro)bot traffic severely hindered the performance of the PURL application and prevented GPO staff from being able to access the system and add/modify PURLs. Under the new hosting solution, the public's and GPO's access is ensured, while still allowing access to traffic, which mostly consisted of link validators.

New PURL Web Address

A new URL of purl.fdlp.gov has been established due to the sunsetting of GPO Access. This new URL ties PURLs to the Federal Depository Library Program. As a result:

• All new PURLs will use the purl.fdlp.gov address.
• Direct requests to purl.fdlp.gov will go to a splash page. The domain is used to resolve to Federal resources.
• All of the 125,000 PURLs that were created using the original OCLC PURL Resolver Software will continue to resolve. No changes will need to be made (at this time) to the existing records that have been made by Federal depository libraries or GPO.

Automated and Enhanced PURL Referral Reporting

As a follow-up from the November 2009 modification of the PURL referral definition and reporting system, an automated PURL referral reporting system was launched on December 1, 2010 that allows Federal depository libraries to view their institution's referrals by individual hostnames and/or IP addresses. The tool also provides a listing of the top fifty (50) referred PURL resources per hostname and/or IP address with:

• The PURL path.
• The full path of the target URL for each PURL.
• The total requests for that individual PURL.
• A search link utilizing the CGP to view cataloging records for the individual PURL.

GPO still releases monthly PURL referral reports; however, these reports include aggregate totals only. Referrals totals strip out bot traffic and focuses on patron requests.

The PURL Referral Reporting Tool is locked down to Federal depository libraries only. Data is current as of the previous day. Historical data is available for twelve months. Tool functionality may be expanded in the future to include greater historical data retention and additional functionality based on funding and community feedback.

Application Monitoring

The PURL application features high availability and failover services. Traffic routing has been optimized to permit PURL resolution as well as search engines and link checkers.

Learn how Federal depository libraries can utilize PURLs at their institution, more about how PURLs operate, and information on the PURL referral reporting tool, by reading Linking to Federal Resources Using Persistent Uniform Resource Locators (PURLs).

The National Plan for Access to U.S. Government Information fulfills the information needs of the public and ensures that Government information is available in perpetuity, when and where it is needed.

Whether in a library or online, our goal is to create, sustain, and enhance programs that ensure Federal Government information is acquired, cataloged, digitized, made findable, preserved for future generations, and disseminated broadly.

In order to achieve this vision, the National Plan sets the groundwork and provides long-term strategies for the success and growth of:

• The Federal Depository Library Program (statutory program)
  The FDLP will continue its evolution to be agile, scalable, flexible, user-centric, and outcomes-based. Extensive partnerships with Federal depository libraries and other Federal agencies and organizations will be key to long-term growth and success. Providing depositories with a rich education program, new services, and flexibilities that will allow for serving diverse communities in the most comprehensive way possible are top goals of the FDLP.


• The Cataloging and Indexing Program (statutory program)
  The C&I Program will continue to enhance its cataloging and metadata creation through the National Bibliographic Records Inventory and its lifecycle management of Government information processes. Acquiring, harvesting, cataloging, discovery tools, and preserving will all play roles in the achievement of a comprehensive Catalog of U.S. Government Publications, both historical and moving forward. Again, partnerships with Federal depository libraries and other Federal agencies and organizations will be key to long-term growth and success.


• The Federal Information Preservation Network (GPO strategic initiative)
  The creation of FIPNet affords the public with guaranteed, long-term access to resources, materials, and expertise from libraries, Government entities, coalitions, and organizations. This program will ensure permanent public access to historic, at-risk publications and resources of significant value to the public. Depository libraries will benefit from the added resources and the sustainability of those resources for their patrons. Based on formal partnerships with GPO, FIPNet partners will provide an added guarantee that our rich history will be accessible for future generations to come. GPO is currently in the process of developing types of FIPNet partnerships, standards, guidelines, and best practices.  More information on FIPNet.

Background

Numerous factors and initiatives provided evidence of the need for a National Plan.

• The FDLP Forecast Study results demonstrated that FDLP libraries need and want services outside the scope of the current FDLP structure.
• A Report by a Panel of the National Academy of Public Administration recommended that GPO take a number of actions to advance its mission and the success and sustainability of its programs.
• There are also a number of external factors that affect evolving library and Government operations.

What began as the “National Plan for the Future of the Federal Depository Library Program,” has evolved into the broader picture of ensuring access to U.S. Government information, of which the FDLP plays one of three parts. For this reason, the National Plan is now formally known as the “National Plan for Access to U.S. Government Information.”

Components of the National Plan

• National Plan for Access to U.S. Government Information (Final version, February 2016)
• FIPNet Information Sheet
• National Plan Information Sheet

1. Introduced FIPNet to the Library of Congress and National Libraries: March 2015
2. Launched the website for the National Plan for Access to U.S. Government Information: April 2015
3. Present the National Plan at the 2015 DLC Virtual Meeting: April 2015
4. Present the National Plan at national library conferences: Summer 2015

Related Resources

National Plan Presentations

• October 2015 DLC Meeting and FDL Conference; Implementing the National Plan: Focusing on Users and Services
• April 2015 Depository Library Council Virtual Meeting: Navigating through Change; National Plan for Access to U.S. Government Information
• April 2015 Depository Library Council Virtual Meeting: Navigating through Change; Process of Change within the FDLP: Open Forum on the National Plan
• December 2014 Depository Library Council Virtual Meeting: Focusing on Service; The National Plan: A Progress Report
• 2014 DLC Meeting & FDL Conference; National Plan for the Future of the FDLP

Back to Top

• Description
• Background
• Components of the National Plan
• Milestones
• Related Resources

In October 2005, the U.S. Government Printing Office (GPO) and Library of Congress Cataloging Distribution Service (CDS) investigated options for electronic dissemination of CDS cataloging publications to Federal depository libraries. At the spring 2005 Depository Library Council Meeting in Albuquerque, GPO staff demonstrated the e-LCSH—a PDF version of the Library of Congress Subject Headings, 28th edition. Federal Depository Library Program (FDLP) partners were then invited to review the e-LCSH and provide feedback by answering a brief survey on the value of LCSH in that format.

The survey results were later shared with CDS for the purpose of assessing the viability of this delivery mechanism and for identifying potential product improvements.

Since the completion of the survey, Library of Congress lists new and changed subject headings on a weekly basis.

In March 2007, representatives from several GPO business units formed a team to conduct a Handles pilot project. Results of the project were to support GPO’s Federal Digital System team on its implementation of Handles.

Handles, like PURLs, are persistent identifiers for Internet resources. Both PURLs and Handles resolve through an intermediary database which can be updated to reflect changes in the location of the internet resource.

At the request of GPO’s Chief Information Officer, Library Services and Content Management (LSCM) staff initiated the creation of Handles in August 2007. After conducting internal (alpha) tests, an external (beta) test was announced at the Depository Library Council Conference in October 2007.

The purpose of the beta test, which ended November 26, 2007, was to gauge the Federal depository library community’s response to Handles and to a range of possible naming conventions. The Handles naming conventions fell into two broad categories – intelligent (containing descriptive characteristics) and non-intelligent (containing no descriptive characteristics and created or assigned automatically).

Beta Test Results

The test, which was conducted from October 15, 2007 through November 26, 2007, revealed the following:

• Access to the requested documents using the Handles was satisfactory.
• There is less interest in the technical aspects of Handles creation and naming than in their utility and other aspects outside the scope of the survey.
• Some libraries would prefer some descriptive content in Handles but not at the risk of jeopardizing their accuracy by increasing the likelihood of typographical errors.
• Non-intelligent naming conventions would be acceptable.

Recommendations

• Handles names should be easily (i.e., machine-) generated and have as much “intelligence” as possible without making them prone to error.
• Handles should be assigned as early in the ingest process as possible.

Requirements

• The library community requires a robust GPO initiative to resolve the problem of broken PURLs, especially in U.S. Catalog of Government Publications (CGP) records.
• In order for PURLs to resolve through the Handle server, Handles need to be created for the approximately 90,000 PURLs that currently exist.

Conclusion

The depository community is more interested in the functionality of persistent names than the specific format utilized to create them.

• Next stages of the PURLs to Handles transition require a coordinated planning effort between LSCM and other areas of GPO.
• Security and support issues involved in continuing the use of PURLs emphasize the need to expedite the transition to Handles.
• Implementation is planned via future FDsys releases.

See where GPO has traveled on our interactive map.

Depository Library Spotlight

Read about previous libraries that were highlighted in the Depository Library Spotlight.

The U.S. Government Printing Office (GPO) is a member of the LOCKSS (Lots of Copies Keep Stuff Safe) Alliance. LOCKSS provides libraries with digital preservation tools and support so they can collect and preserve their own copies of authorized electronic content.

GPO’s work in this area is in keeping with our mission to provide permanent public access to official Federal Government publications in print and electronic formats through the Federal Depository Library Program (FDLP). The FDLP was established by Congress to ensure that the American public has access to its Government’s information. Since 1813, depository libraries have safeguarded the public’s right to know by collecting, organizing, maintaining, preserving, and assisting users with information from the Federal Government.

In a July 2010 interview in Library Journal (LJ), Ric Davis, GPO's acting Superintendent of Documents, told LJ that libraries working with GPO have expressed interest in having FDsys established as a trusted digital repository as well as the possibility of having their own local copy of GPO content.

"The foundation of the Federal Depository Library Program was built on a distributed model, with tangible publications being held throughout the country," Davis said. "We're looking at how we can continue that activity in the electronic world."

GPO's participation in the LOCKSS alliance is a step toward fulfilling the interests that have been expressed by the Federal depository library community.

What is LOCKSS?

LOCKSS (for"Lots of Copies Keep Stuff Safe") is open source software that provides institutions with a way to collect, store, and preserve access to their own, local copy of content. LOCKSS was developed by Stanford University, and it is currently maintained by the Stanford University LOCKSS Program Management Office with support from the LOCKSS Alliance. LOCKSS runs on standard desktop hardware and requires minimal technical administration. Once installed, the LOCKSS software converts a personal computer into a digital preservation box that creates low-cost, persistent, accessible copies of e-journal content as it is published. The accuracy and completeness of content stored in a LOCKSS box is assured through a robust and secure, peer-to-peer polling and reputation system. A LOCKSS box performs the following four functions:

• It collects newly published content from the target e-journals using a Web crawler similar to those used by search engines.
• It continually compares the content it has collected with the same content collected by other boxes, and repairs any differences.
• It acts as a Web proxy or cache, providing browsers in the institution's community with access to the publisher's content or the preserved content as appropriate.
• It provides a Web-based administrative interface that allows the institution staff to target new journals for preservation, monitor the state of the journals being preserved, and control access to the preserved journals.

Collecting

Before LOCKSS boxes can preserve a journal, two things have to happen:

• The publisher has to give permission for the LOCKSS system to collect and preserve the journal. They do this by adding a page to the journal's Web site containing a permission statement and links to the issues of the journal as they are published.
• The LOCKSS box has to know where to find this page, how far to follow the chains of Web links so that it doesn't crawl off the edge of the journal and try to collect the whole Web, some bibliographic information, and so on. In order to add new publishing platforms, the LOCKSS system provides a fill-in-the-blanks tool that a librarian or administrator can use to collect this information and test that it is correct. The information is then saved in a file (the LOCKSS plug-in) and added to the publisher's Web site or to some other plug-in repository, so that it is available to all LOCKSS systems.

Preserving and Auditing

The LOCKSS boxes at libraries around the world use the Internet to audit, continually but very slowly, the content they are preserving. At intervals boxes take part in polls, voting on the digest of some part of the content they have in common. If the content in one box is damaged or incomplete that box will lose the poll, and it can repair the content from other boxes. This cooperation between the boxes avoids the need to back them up individually. It also provides unambiguous reassurance that the system is performing its function and that the correct content will be available to readers when they try to access it. The more organizations that preserve given content, the stronger the guarantee they each get of continued access.

Providing Access

LOCKSS boxes provide transparent access to the content they preserve. Institutions often run Web proxies, to allow off-campus users to access their journal subscriptions, and Web caches, to reduce the bandwidth cost of providing Web access to their community. Their LOCKSS box integrates with these systems, intercepting requests from the community's browsers to the journals being preserved. When a request for a page from a preserved journal arrives, it is first forwarded to the publisher. If the publisher returns content, that is what the browser gets. Otherwise the browser gets the preserved copy.

Administering

Staff administer their LOCKSS box via a Web user interface. It allows for targeting the appliance to preserve new journals, monitoring the preservation of existing journals, controlling access to the box and other functions.

Additional information about LOCKSS is available from the Stanford University LOCKSS Web site.

Pilot Project

GPO received numerous requests from research institutions, universities, depository libraries, and other Federal Government agencies to investigate using LOCKSS as a means to manage, disseminate, and preserve access to Web-based Federal Government e-journals that are within the scope of the FDLP and the IES. As a result, GPO conducted a 12 month pilot to make Federal Government e-journals available to select pilot libraries that are operating LOCKSS boxes. The following provide a summary of GPO's evaluation of the pilot project.

 pdfExecutive Summary (40 KB)

 pdfReport (282 KB)