What a serials publisher needs to know to be a serials cataloger

This article illustrates why publishers need to care about library catalog records. Back ground information on serials cataloging practices and the International Standard Serial Number (ISSN) are provided to foster a common understanding of serials cataloging across the serials community. The recently implemented CONSER standard record is discussed and elements of the record are identified that could be provided by pub lishers and others in the serials community.

As libraries continuously investigate ways to streamline and simplify the cataloging process, one consideration is duplication of effort and information. It is useful to consider where serials metadata is being unnecessarily duplicated between the various players in the serials supply chain (publisher, vendor, public access management service (PAMS) and library) and what standards and tools we can use to reduce this duplication. Recent work has been done involving vendors and PAMS distributing publisher holdings (coverage) information to libraries in the UK 1 and the US. 2 The CONSER program in the US and, more recently, the SUNCAT project in the UK, are successful examples of libraries working cooperatively to create, maintain and distribute serial catalog records, but the data for these records is created by individual libraries and then distributed to other libraries (and vendors/PAMS who then redistribute to libraries). This paper investigates ways that bibliographic data produced by the publisher can be used by the library cataloging community.
There are a few examples of publishers serving as a source for bibliographic record information. One is a CONSER project focused on publishers and vendors of aggregated full-text databases with the idea that the database producers could also create library catalog records for the individual serials within the databases. 3 Another example of catalog record production from publisher-supplied metadata is associated with the assignment of the International Standard Serial Number (ISSN). Several ISSN Centers (including the UK and Canadian centers) have online application forms 4,5 that publishers can use to apply for an ISSN. In some cases, submission of the online application results in the creation of a library catalog record which can then be edited by a cataloger to create a serial catalog record that meets all the current requirements for national-level cataloging programs.
These efforts are good starting points but it may be possible for the library cataloging community to take advantage of more data produced by serials publishers. Until recently, one barrier has been the record format used by libraries. The MARC (MAchine-Readable Cataloging) format was developed in the early 1970s at the Library of Congress and currently is the primary record format used by library systems. Although most library systems are MARC-based, catalogers are no longer required to work in native MARC as metadata standards -crosswalks and interfaces -exist that can convert non-MARC data into a MARC record which can be used in existing library systems. So the problem may not be the metadata schema used but instead the identification (ISSN) and content standards (e.g., cataloging rules) that catalogers use to create the data. What follows is one cataloger's perspective on what a serials publisher should know about serials cataloging, ISSN, and the newly developed access level record for serials to better prepare publishers to consider supporting the serials cataloging function.

What a serials publisher needs to know to be a serials cataloger
This article illustrates why publishers need to care about library catalog records. Background information on serials cataloging practices and the International Standard Serial Number (ISSN) are provided to foster a common understanding of serials cataloging across the serials community. The recently implemented CONSER standard record is discussed and elements of the record are identified that could be provided by publishers and others in the serials community.

STEVE SHADLE Serials Access Librarian University of Washington Libraries, Seattle
But this discussion (and the title of this article) begs the question: Why should a publisher care about library catalog records? Libraries are an important (if not critical) customer base for serial publishers. 6 Studies have shown that adding e-serial records to the library catalog dramatically increases database and e-serial usage. 7 With the advent of electronic resource management (ERM) systems that ingest usage from various providers (based on COUNTER and SUSHI standards), libraries will be able to better identify usage and cost-per-use for electronic resources. Publishers concerned about the usage of their journals should do all they can to maximize usage of their publications, and the provision of library catalog records is one way of doing that.

The serial title and how it changes
It is in the nature of any continuing resource that its title will likely change over time. The serial title is one of the primary identification elements used in citation practice. (Some citation standards will not include publisher or place of publication, but all will include journal title). This is changing to some degree on the web where citation styles may include a URL or uniform resource identifier (such as a DOI), but even with these newer standards, the serial title is still incredibly important in citation practice.
In the library catalog, there are three possible cataloging treatments to describe serial title changes, and the cataloging community has followed all three at one time or other.

Earliest entry
One is to enter a serial in a catalog or listing under its earliest title, make notes about later titles, and make references from later titles to the complete catalog record listed under the earliest title. This practice was used in the United States through the 19th century until the 1908 adoption of the ALA cataloging code. 8 In the context of the library catalog of the time (the printed or manuscript book catalog) this treatment made sense as, firstly, duplicating a complete catalog record under every title variant was time-and space-consuming, and, secondly, maintenance was easier because handwritten or typewritten notes (in the case of a book catalog with removable pages) could be added to the existing record (rather than having to rewrite a complete entry elsewhere in the catalog). References would guide the user to the complete catalog record under the earliest title.

Latest entry
With the ALA cataloging code came the advent of 'latest entry' cataloging where the complete catalog record (typically transcribed on a catalog card in a card catalog) was filed under the latest title. Before the era of affordable reprography, it still made economic sense to have only one complete catalog record in the library catalog, but that record could now be edited and re-filed more easily. The complete catalog record would be found under the latest/current title. However, the result of having the complete history of a serial on a 3x5 card was a catalog record that often would not fit on one, or even two, cards. The resulting catalog records were more complex to file and more difficult to read and interpret.

Successive entry
The first edition of the Anglo-American Cataloguing Rules (AACR, 1967) introduced the concept of 'successive entry.' The intent of the new code was "to prepare an entry that will stand the longest time and will permit the making of necessary changes with the minimum of modification." 9 One of the principles of the new code was that a separate (successive) entry is made when a serial title changes. However, minor variations in title which "do not affect the location of the title in an alphabetical file or conceal the identification of the parts" 9 were simply noted on the catalog record. In the context of the card catalog, the obvious benefit was the when a title changed, the main entry card (with the complete catalog record) did not need to be recreated based on the latest title. Instead, the catalog record could be closed, with a note added about the new title. The new title would then have a separate catalog record (which was left more or less intact until the title changed again). The new practice resulted in less catalog maintenance (no need for reference cards as each significant title variant was basically its own catalog record) and more concise records that were easier to understand (as they described a smaller set of issues). However, catalog users who were interested in the history of a serial title were required to thread their way through several locations in the catalog to identify the information associated with each title.

The ISSN, title changes and harmonization
All this is interesting (for some catalogers anyway) … so why the history lesson? The International Standard Serial Number was developed in the early 1970s by the International Organization for Standardization (ISO) because of the need for a brief, unique and unambiguous identifier for serial publications. Successive entry cataloging was the serials entry convention used at the time the ISSN standard (ISO 3297) was developed, and ISO 3297 incorporated the concept of successive entry cataloging by requiring that a new ISSN be assigned when a serial changes its title.
So who makes this assignment? Responsibility for ISSN assignment rests with the ISSN Network, a UNESCO-designated program which maintains an ISSN International Centre in Paris that coordinates the ISSN assignments made by 81 national ISSN centers. Unlike the International Standard Book Number (ISBN) where the assignment is made by the publisher (based on a block of ISBN the publisher purchases from the ISBN Agency), final authority for ISSN assignment rests with the national and international ISSN centers. Serial publishers should have regular contact with their corresponding ISSN center (the ISSN UK Centre 10 for publishers located in the UK, the National Serials Data Program 11 for US publishers). If a serial title is transferred from one publisher to another, the ISSN remains the same (unless the title also changes). The national ISSN center responsible for the assignment may change (such as in the case where a title published by a UK publisher is transferred to a US publisher) but unless the title itself changes, the ISSN will remain the same.
However, not every change in title will result in a new ISSN being assigned. With the advent of successive entry cataloging, catalogers quickly learned that the previously cited rule ('location of a title in an alphabetic file') was not sufficient to be able to apply consistently (as your alphabetic file is not the same as my alphabetic file and how does one file abbreviations, acronyms and punctuation in any case? Is three cards away still close enough to be the same location?). The second edition of the Anglo-American Cataloging Rules (AACR2) published in 1978 sought to remedy this ambiguity by providing more specific rules. In a nutshell, the rules stated that any change of title in the first five words (not counting initial article) that affected important words ("nouns, proper names or initials standing for proper adjectives, etc.") was considered a change that would require a new entry. Thus Fishery Report to Fisheries Report or Open House to Openhouse were title changes that would require a new entry and a new ISSN. However, this was not seen as a tenable solution (there were still too many successive records for unintended title changes) and the Library of Congress began issuing rule interpretations so that more of these unintentional changes would be considered title variants rather than title changes. Over the following 20 years, catalogers got better at identifying the kinds of title changes that were not intentional or significant and these were incorporated into later revisions of AACR2. In 2000 and 2001, representatives from the Joint Steering Committee for Revision of Anglo-American Cataloguing Rules (JSC), the ISSN Network and the International Federation of Library Associations (IFLA) met with the goal of harmonizing serials cataloging title change rules with an eye towards minimizing the creation of new serials catalog records and ISSNs. The resulting rules have attempted to minimize the number of unnecessary records and ISSN assignments while at the same time recognizing the fact that serials do intentionally change their titles and that the resulting standards need to fit within the context of successive entry treatment.
In summary, the current rule is to consider any change in the first five words (not counting an initial article) to be a major title change (i.e., requiring a new record/ISSN) unless: Because of the widespread use of the ISSN and the fact that the ISSN is so directly tied to the serial title, in an ideal world the publisher would know to consult with their ISSN center when considering a title change (or ideally would know whether a title change they are considering is one that would result in a new ISSN).

Bibliographic sources, or, "But that's not the title!"
One consideration that the publisher may not be aware of is that the physical location of title presentations on an issue is crucially important to catalogers.

The access level record for serials
The CONSER database (along with the ISSN Register and SUNCAT database) is one of the most useful sources of serial catalog records for libraries worldwide. The 53 libraries, agencies and services that contribute to the CONSER database follow a set of fairly extensive practices that are spelled out in the CONSER Catalog Manual and CONSER Editing Guide. Following the success of the Library of Congress' development and testing of an 'access' level record for electronic resources, 12 a working group within CONSER was charged with using a similar process to develop and test an access level record for serials. 13 One hope was that with cataloging simplification, some elements in the CONSER record could be provided by serial publishers (or others in the serials life cycle) and that CONSER catalogers could then build on the publisher-provided elements to create a CONSER standard record. Although there was not unanimous agreement on the conclusions reached, the Program for Cooperative Cataloging (the umbrella organization for several US cooperative cataloging programs) did endorse the report and (working with the Library of Congress' Cataloging Policy and Support Office) agreed to a partial implementation on 1 February 2007, with full implementation later in 2007.
The CONSER standard record will apply to serials in all formats (including online) and will reduce serials cataloging costs by requiring in serial records only those elements that are necessary to meet important user tasks. Cost savings and user benefits will be realized by eliminating redundant information in the catalog record and by recording elements in a way that is more straightforward for the cataloger to provide and easier for the end-user to understand. The record is intended to be a 'floor', to which additional elements can be added if such elements are essential to meeting user needs for a specific resource or to meet the needs of a particular institution.
What follows are two annotated summaries of the required data elements. The first set of elements are those that the publisher would have knowledge of and can contribute to a catalog record. The second set of elements are those that can be contributed by serial catalogers. Appendix A provides the complete list of elements as specified in the report. ■ additional coded information (e.g. beginning/ ending years, country, language and format of publication).

Summary of cataloger provided elements
In addition to those elements provided by the publisher, those which catalogers would continue to provide include: ■ additional control or identification numbers ■ classification numbers (e.g., Dewey Decimal, Universal Decimal or Library of Congress Classification) ■ main entry (No, not the front door). In singlefile systems, this is the primary access point chosen for filing ■ subject headings and/or thesaurus terms (and cross-references for those subject terms) ■ organization names (and cross-references for those names) ■ other editions (translations, language, format, regional, audience) or supplements ■ series access point (a cataloger-created 'official' form of series title and series cross-references) ■ additional MARC record elements that support record processing.
As can be seen from this list, establishing references (so that users can search under different forms of the same name [e.g. UKSG, UK Serials Group, or United Kingdom Serials Group] and retrieve the same set of documents) is an important part of the cataloger's job. Having publishers create brief bibliographic records would provide the cataloger with more time to spend on elements like references (also known as 'authority work') and subject access, elements which greatly assist potential users. Catalogers might even have time to increase serial catalog record production, thus providing intellectual access to more serial titles.

Electronic serials
There are some cataloging issues that are specific to electronic serials that should also be considered. The primary one is the preservation of title history and integrity. When a serial has undergone several changes of title (and ISSN), it can be convenient for the publisher to maintain only one website for the entire run of that serial. Typically in this situation, the only title and ISSN presented are the current title and ISSN. This is similar to latest entry cataloging and publishers do researchers and librarians no justice in following this practice. Scholarship is based on the practice of citationbased research and users with a citation for an earlier title may not recognize that the article they are looking for is listed under a later title. Abstracting and indexing database entries will likely include the ISSN and title of the journal at the time the article was published. PDF versions of articles will also present the journal title as it appeared when the print version was originally published. This discrepancy between online and print version presentations may cause problems with link resolution (if an earlier ISSN is not accounted for) and can be confusing for users when faced with a presentation of title that does not seem to be the same title as that appearing in a citation or reference 14 . Publishers can help ameliorate some of this confusion by consciously recognizing title changes and earlier ISSNs (either by having separate, linked websites for each successive title or at least some acknowledgement within the website of each earlier title and ISSN).
Currently, there are no other identifiers used in the citation of electronic serial content which are as effective as the ISSN. Citation practice which includes a URL is becoming more commonplace but research has shown that URL linkrot is commonplace in bibliographic citations. 15 The publisher-assigned Digital Object Identifier (DOI) has great potential for being a more stable identifier, but problems still exist in CrossRef linking 16 that prevent it from being the one, single identifier associated with a bibliographic citation. Until that day, publishers should invest time and energy in understanding the ISSN and how it is used in various functions and systems.

Next steps
How can librarians get publishers to help us with our work as serials catalogers? A first step is just to maintain contact with your ISSN center. Consequent with that, if more ISSN centers were to develop online submission forms that created real (but brief) MARC records, serials catalogers would be one step closer to being out of a job! Now that the technology and standards exist to crosswalk publisher data into a format usable by libraries, would publishers consider going into the business of providing catalog records to their customers? Do publishers consider the serial catalog record as strictly a library 'thing'? Or would it be beneficial for them to understand why and how the catalog record and cataloging standards support intellectual access to their publications? The Serials Cataloging Cooperative Training Program (SCCTP) 17 has been a successful model in the US for serials cataloging continuing education. If there is sufficient interest from the commercial sector, perhaps similar education (tailored to publishers and vendors) could be offered in the UK.

Conclusion
Historically, the information systems used by various players in the serials life cycle have not been integrated or even well-coordinated. Catalogers, vendors and publishers worked within their own internal systems (with their own data formats and content standards) with little connection to each other. With the increasing development of data standards and crosswalks, there has been more transfer and recycling of data between various systems and functions. However, one area that has remained relatively isolated is the creation and maintenance of library catalog records. Publishers, vendors, PAMS and libraries have the potential to work together to streamline the production of metadata that provides access to and supports identification of the serial literature. It is this author's hope that understanding the identification (ISSN) and content standards (cataloging rules) within which the serials cataloging community operates will help all in the serials community work towards better, more interoperable systems and services. For a link to the table of contents for the issue of Serials in which this article first appeared, click here: http://serials.uksg.org/openurl.asp?genre=issue&issn=0953-0460&volume=20&issue=1