CONSER proposes that participants explore their options for contributing data using various methods (direct editing on OCLC, transfer of information from a serials control system, etc.) We also recognize that library systems have widely varying capabilities in storing this data, in outputting it to other systems or utilities, and in using it for local holdings recording and serials management. The CONSER Patterns and Holdings Project is an experiment, in which we hope that libraries will become involved to the level of their capability and interest in seeding a database with data that can be appraised for its quality and ease of use, in hopes of building an archive that all can benefit from.
With the exception of batch-loaded data, data should be supplied based on a piece in hand or other information determined to be reliable.
All 891 fields input directly to OCLC (as opposed to batch-loaded data) should normally be based on current issues of the serial. Every attempt should be made to add new patterns as soon as possible once subsequent changes occur that require a new pattern.
All newly input data should be as complete as possible. Each statement must consist of paired fields. Only fields for basic units (rather than supplements and indexes) are mandatory. The first of the pair contains captions and patterns (891 $9 853). The second contains enumeration and chronology (891 $9 863). Each caption/pattern statement must have at least the captions portion of the statement (appropriate subfields for the captions corresponding to the piece in hand, $a-g and $i-n.). Each enumeration/chronology statement must have the same sequence of subfields, since the actual data is correlated to the captions and patterns. Beyond these essentials, input of pattern data ($p-x) is strongly encouraged, since pattern subfields are the ones that will be used by libraries to control the receipt of the publication. Participants whose system outputs pattern information automatically should be aware that some values, e.g. $w (frequency) $x (calendar change), and $y (regularity pattern) may not be automatically generated and should be supplied manually if necessary. Data for alternative enumerations ($g-h) and chronologies ($m-n) is optional. Data for supplements and indexes is to be input only in special cases. For further information, see the following sections.
Participants are encouraged to complete data in patterns found in CONSER records when the information is readily available.
Data included in 891 fields should comply as closely as possible with the MARC 21 Format for Holdings Data (see the Concise version at http://www.loc.gov/marc/holdings/echdhome.html)and ANSI/NISO Z39.71-1999 Holdings Statements for Bibliographic Items (available for free download at www.niso.org/standards/index.html (external link)). Data that is supplied to circumvent local system limitations or that reflects non-standard use of the MFHD or ANSI standard by a system is discouraged. However, it is up to the institutional liaison to determine when to contribute pattern data to CONSER records, based on a knowledge of the MFHD and their system's capabilities and limitations.
Pattern data that has been batch-loaded from various institutions may or may not be current or complete and should be updated or completed as necessary by participants when they encounter the record. See also Appendix C, Harvard Load records.
Participant-input data may be completed if incomplete. Changes to the data (when there has been no change in the serial) should be made only when the data is not in compliance with the MFHD or ANSI standard. Do not change pattern data nor add additional fields to resolve local system needs.
Changes in either the captions or the pattern can necessitate input of a new pair of fields to indicate the changes and the issue in which the changes were noted. Minor changes may enable simple modifications to the pattern. Refer to Section 5.3, When to input a new caption/pattern, for further information.
Because the pattern of publication relates to information ordinarily given in the bibliographic record:
frequency, which is noted in the 310 (Current frequency) and 321 (Former frequency) fields as well as coded in 008/18 (in OCLC, Fixed Field code Freq)
regularity, which is coded in 008/19 (in OCLC, Fixed Field code Regl)
numbering, which is given in the 362 field, with any needed explanations of "peculiarities" given in the 515 field, a change in the pattern commonly will necessitate a change in one or more of those fields in the bibliographic record. Changes to 853 subfields $a-h, $w, and $y are the important changes to watch for. Refer to section 7 for complete instructions.
The OCLC 891 field has been specially designed for the CONSER Patterns and Holdings Project to "stand in" for holdings fields. The substitution is made by including a $9 subfield for the destination tag. As explained in the Project, the regular holdings fields were already used by OCLC for local data; but they offered the 891 as an alternative. The Project will employ:
a) 891 with subfield $9 853 routinely to store the captions and publication pattern information for the basic units (principal numbering sequence). See Section 5.
b) 891 with subfield $9 854 and 855 rarely to store the captions and publication pattern information for supplements and indexes if those are determined to be regularly issued and of special importance for the pattern; but only when there is room in the record for this information. See Section 5.
c) 891 with subfield $9 863 routinely to store the enumeration and chronology of the first bibliographic unit issued or known to have been published with that pattern. (See the detailed instructions in Section 6.) The data in this field should correspond to the actual enumeration and chronology of the serial as published, rather than only to parts held, whenever possible. In cases where the library does not know the first issue corresponding to the publication captions and pattern, or prefers to establish the pattern from an issue in hand, the issue may be given in angle brackets. See Section 6.
d) 891 with subfield $9 864 and 865 rarely to store the enumeration and chronology of supplements and indexes, only when these are determined to be regularly published and have special significance for the pattern. See Section 6.
Record length problems are alleviated in OCLC Connexion. The number of variable fields is now unlimited (previously 50), but no single field should exceed 4,095 characters. The length of a bibliographic record is more than 20,000 characters (previously 4,096 characters). The Initiative's Task Group on Long Term Storage and Task Group on the Uses of a Publication History Record will aim at solving the question of storage of the 891 fields, reformatted as a Publication History Record.
Though the MFHD allows input of chronology data (months and seasons) in either natural language or in the numeric equivalents 1-12, 21-24, participants are asked to input these values in numeric form, rather than in natural language. Numeric form is necessary for triggering the prediction algorithms in serials control systems. This procedure is followed, however, only for data within the chronology subfields $i-m, not when chronology is placed in enumeration subfields (because the serial lacks other enumeration). See Section 220.127.116.11. and corresponding example.
OCLC supplies a macro called “Generate891” for adding publication patterns with Connexion client upgrades.
Field 891 is an OCLC-defined field. It contains holdings data, specifically the content prescribed for 853-855 and 863-865 in the MARC Format for Holdings Data. The destination tagging is specified by means of a subfield $9, which is the first subfield within the field. In the CONSER project, all data content is composed of captions and publication pattern for base volumes, supplements, or indexes (853, 854, and 855), and corresponding enumeration and chronology (863, 864, and 865). Therefore, the $9 subfield will always carry one of these values.
Because the CONSER project is dedicated to summary information at the national level, some of the values given for the holdings fields in the MARC Format for Holdings Data are not used. Though most are listed, not all of them are explained in full. We encourage all participants to refer to the Format for further information on field content and encoding. [The USMARC Concise Format for Holdings Data is available on the Web.]
891 _ _ $9 853 Captions and pattern for base volumes
891 _ _ $9 854 Captions and pattern for supplements
891 _ _ $9 855 Captions and pattern for indexes
891 _ _ $9 863 Enumeration/chronology for base volumes
891 _ _ $9 864 Enumeration/chronology for supplements
891 _ _ $9 865 Enumeration/chronology for indexes
The use of the 85X field and 86X field in combination will show the correspondence of a set of captions and patterns to the particular parts of the publication designated by them. In the case of a current pattern, a subfield $3 can be added to 853 to give the last issue known to bear this pattern.
Questions? Contact the Project at: CONSERHOLD-L@cornell.edu
rev. dih 2/26/01