Discussion Paper 2001-DP06

DATE: May 21, 2001
REVISED:

NAME: Coding Series Numbering in the MARC 21 Bibliographic and Authority Formats

SOURCE: PCC Working Group on Series Numbering

SUMMARY: This paper describes a situation that suggests a possible change to subfield $v in the 4XX and 8XX fields of the MARC 21 bibliographic format, and/or a change to the 642 field of the MARC 21 authority format for proper sorting of series headings. The paper suggests either a change to coding practice using existing MARC content designation, or a change to MARC content designation.

KEYWORDS: Fields 4XX (BD); Fields 8XX (BD); Field 642 (AD); Series added entry fields (BD); Series numbering example field (AD); Series statements fields (BD); Subfield $v, in 4XX fields (BD); Subfield $v, in 8XX fields (BD)

RELATED:

STATUS/COMMENTS

05/21/01 - Made available to the MARC 21 community for discussion.

6/18/01 - Results of the MARC Advisory Committee discussion - Although the participants considered the coding solutions presented in the discussion paper as viable ones, they decided that the problems associated with sorting series headings were largely system-related. It would be useful to develop specifications for an algorithm to correctly sort series headings to be shared among library system vendors to facilitate better sorting. This issue will be initiated on the MARC electronic discussion list.

Discussion Paper No. 2001-DP06: Coding Series Numbering in the MARC 21 Bibliographic and Authority Formats

1 BACKGROUND

In 1999, BIBCO, prompted by concern on the part of many PCC members about the display of series headings in library catalogs, appointed a Working Group on Series Numbering (WGS) to consider changes to the presentation of series information in bibliographic and authority records and related matters. The Working Group presented its final report in the same year.

In conducting a survey of these issues, the WGS asked a broad range of PCC and other libraries if an inconsistent use of the caption/designation in series subfield $v causes a problem for public service librarians, and gave an example of such inconsistent use in question three of its survey instrument. Approximately 68% of the respondents said that the inconsistent display causes a problem for the OPAC user. In addition, the WGS asked how many library systems are able to sort series on the volume number rather than on the caption/designation that can appear in subfield $v (question six). A large majority (86%) said that their systems do not have such a capability. The WGS asked librarians if they think that library system vendors should help remedy the problem with series sorting (question thirteen); 90% answered "Yes" to the question.

The WGS thus clearly established that the failure of the current generation of library systems to sort series headings in numerical order hinders the proper display of series information in public catalogs, and demonstrated an interest on the part of its respondents for a solution to the problem. Among its recommendations, the impetus for the present discussion paper, was the following:

4. MARBI and vendors should work toward developing a mechanism which supports disregarding the designation in the series $v in its sort of the series.

This discussion paper is a direct outgrowth of the work of the Working Group. The WGS is asking that some mechanism be provided in the MARC formats that will enable designers of library systems to provide displays of headings for series, arranged by their numbering. No matter what the outcome of this discussion paper, the WGS hopes that MARBI will join it in encouraging vendors to sort bibliographic series headings in numerical order.

2 DISCUSSION

Where they provide any useful guidance at all, standard sets of rules for sorting entries in library catalogs imply that numbered series headings should be arranged according to the numbering considered as a value and not as a string of characters. However, some library systems that attempt to provide access to the numbering present in subfield $v series headings have not followed these guidelines; they typically provide such access by arranging information found in subfield in alphabetic order, rather than in the numeric order that is more desirable. This produces online displays similar to the following:

Series heading ; v. 1
Series heading ; v. 10
Series heading ; v. 11
Series heading ; v. 2
Series heading ; v. 20
Series heading ; v. 21
Series heading ; v. 3
Series heading ; v. 30
Series heading ; v. 4

Such a display is confusing to public catalog users and unhelpful to public services staff, especially when many bibliographic records contain the series heading. This sort makes it very easy for users to miss the volume of interest. The problem becomes even more severe if there is any variation in the caption that accompanies the numbering. Although the series authority record prescribes the proper caption to use, records often arrive with variant captions, and this variation is not always caught. Failure to properly apply the data content rules produces displays similar to the following:

Series heading ; no. 7
Series heading ; no. 8
Series heading ; v. 1
Series heading ; v. 2-3
Series heading ; v. 9
Series heading ; vol. 4

Finally, the series numbering may contain more than one discrete piece of numeric information, further complicating the construction of meaningful displays:

Series heading ; ser. 74, no. 1-3

In the MARC 21 bibliographic format, subfield $v is defined in the 4XX (series statements) and 8XX (series added entries) fields for the "volume/sequential designation" (i.e., the "numeric or other designation" spoken of in AACR2) of a series heading. In the MARC 21 authorities format, subfield $a of the 642 field contains an example of the caption and numbering as a pattern for data in subfield $v of bibliographic 4XX and 8XX fields. In neither format is a mechanism provided for indicating which parts of the series numbering subfield should be arranged alphabetically, which parts should be arranged numerically, and which parts (if any) should be completely ignored. Since most systems do not try to sort subfield $v numbers numerically and catalogers do not provide consistent data, there is perhaps a format change that would assist the former and neutralize the latter.

3 POSSIBLE SOLUTIONS

Subfield $v in series headings presents the system designer with two problems: it contains text that needs to be disregarded (in most cases), and it contains other text that needs to be filed as a value rather than as a string of characters. The first problem might be solved with some change to the MARC bibliographic format, and the application of that new coding by those creating and maintaining bibliographic records. This problem is the main subject of the remainder of the present discussion paper. But it is likely that a unified solution to both problems, and an achievement of a more reasonable sorting of series headings, cannot be achieved without changes to library systems independent of any change to coding. It is even possible that, given the variety of information that may constitute a series heading, no solution that produces an acceptable display for all series in all cases will ever be found.

Several possible solutions to the problem of text in subfield $v that should be disregarded immediately suggest themselves. These solutions may have substantial implications for existing data; retrospective conversion may be the only means to achieve the goal of proper sorting of series headings. Any change to existing practice approved as a result of this discussion paper may have its principal effect on catalogers, utilities and providers of library information systems.

These solutions include:

  1. In the bibliographic record, mark text to be disregarded with the recently-defined (but not yet implemented) control characters (hex �' and �'). In these examples, the non-sorting begin and end characters are represented by braces.
    $v {v. }1
    Interpretation: The subfield contains only the numeric value "one."

    $v {ser. }74{, no. }1-3
    Interpretation: The subfield contains the numeric value "seventy four" followed by the value "one" followed by the value "three."
  2. Define a subfield in the bibliographic format for the non-sorting portions of series numbering; because the numbering may consist of more than one number, subfield $v would need to be redefined as repeatable. These examples use subfield $? for the non-sorting characters subfield.
    $? v. $v 1
    $? ser. $v 74 $? no. $v 1-3
  3. Define a new bibliographic subfield to contain a normalized sorting form of the series numbering. This subfield (whose contents would have to be specified in some detail) would be similar in function, if not appearance, to the field 456 and similar fields defined in the MAB2 format (bibliographic format used in Germany). Library systems would use the series numbering in this subfield instead of the display version contained in subfield $v when sorting series headings for displays. In the following examples, all for the same series numbering, subfield $? contains the sorting version of the series numbering; different conventions for recording information in this subfield are illustrated, simply to show that various conventions might be possible.
    Series heading; $v v. 17 $? 17

    Series heading; $v v. 17 $? 00017

    Series heading; $v v. 17 $? 517
    Interpretation: The series numbering "17" should be normalized into a field 5 digits wide.
  4. Regardless of the mechanism chosen for recording information in bibliographic records, a parallel mechanism might be applied to the 642 field (series numbering example) in the MARC authority format. It is even possible that in many cases proper coding in the 642 field of authority records could be employed by library systems as a pattern for use in the interpretation of subfield $v in older bibliographic headings, without requiring the application of revised coding to existing bibliographic records.
    $a 2 $5 IEN
    Interpretation: Subfield $v in bibliographic records contains only numeric values; disregard other information found in the subfield.

    $a {no.} 13 $5 IEN
    Interpretation: Subfield $v in bibliographic records contains only numeric values, which should normally be preceded by the caption "no."; disregard information in the subfield other than the numeric value.

    $a {no.} FDA 79-2108 $5 IEN
    Interpretation: Subfield $v in bibliographic records contains a mixture of alphabetic and numeric characters, which should normally be preceded by the caption "no."; disregard such a caption when present, then sort alphabetically on alpha characters and numerically on numbers.
  5. Instead of making any change to MARC content designation, instruct system designers to disregard alphabetic characters when present together with numeric characters in subfield $v. Although this would work well enough for the simple examples shown elsewhere in this paper, it would not work in other cases in which at least some of the alphabetic characters in subfield $v should be considered in sorting. These examples will be challenging for other solutions since non-numeric data needs to be considered in the sorting.
    $v no. (FDA) 79-2108 EV
    $v MSHA IG 8
    $v no. 97-A-97-D
    $v FS-99-168
    $v EDO-RC-00-6

4 QUESTIONS

  1. What mechanism best allows the proper sorting of series entries in catalogs?

  2. What are the ramifications for retrospective records and database cleanup? How much of a problem will be created by files that contain records coded according to the new convention, and other records?

  3. Could revised coding of the 642 field in authority records be used to modify existing bibliographic records? Or instead of any change to coding in bibliographic records, could systems use such revised authority coding to parse the subfield $v and index it correctly? Since this would depend on correct data patterns in the bibliographic record, could the authority 642 data be used to validate the bibliographic record $v strings?

  4. Should subfield $v in bibliographic records alone be coded to identify sorting and non-sorting information, or only the 642 field in authority records, or both?

  5. What can we do to encourage the bibliographic utilities and integrated library system vendors to do a better job of sorting series through the entire series numbering?

Go to:

Library of Congress Library of Congress
Library of Congress Help Desk (10/12/01)