DATE: May 24, 2001
NAME: Incrementing intervals in publication patterns in the MARC 21 Holdings Format
SOURCE: CONSER Task Force on Publication Patterns and Holdings
SUMMARY: This paper considers methods to code for titles with multiple basic components that sequence their enumeration with skipped numbers, such as the practice of using either even or odd numbers to identify serial issues.
KEYWORDS: Field 853-855 (HD); Subfield $s, in Field 853-855 (HD); Publication pattern (HD); Incrementing intervals (HD)
5/24/2001 - Made available to the MARC community for discussion.
6/16/01 - Results of the MARC Advisory Committee discussion - Participants expressed interest in providing a method for coding titles with multiple basic components that sequence enumeration with skipped numbers. However, many members were concerned that system vendors may have difficulty providing the complex algorithms required for such serial prediction. Participants favored using subfield $y to include a new code for enumeration regularity. The group asked that other ideas and possible solutions be posted to the MARC electronic discussion list for consideration. A proposal may be written reflecting these decisions for the midwinter meeting.
The current publication pattern fields defined to predict serial issue descriptions do not recognize the situation where different components of the same serial, each bearing a distinctive title, have intermeshing or inter-sequenced serial numbering. The numbering of such issues must be synchronized between components but does occur at regular, predictable intervals. One example is a title with dual basic bibliographic units that are separately titled. It generates issues in the following type of pattern.
Section A: v.1, no. 1, v.1, no. 3, v.1, no. 5, v.1, no. 7, v.1, no. 9, v.1, no. 11
Section B: v.1, no. 2, v.1, no. 4, v.1, no. 6, v.1, no. 8, v.1, no. 10, v.1, no. 12
Intermeshing the two sequences creates a single sequence of numbers 1-12. The issues come out in alternating months or weeks within each component. Current ILS systems force prediction such that a title, that generates 6 issues per section over a year period must be expressed as follows:
853: $a v. $b no. $u 6 $v r $i (year)
$w 06 $yom 02,04,06,08,10,12
863 $a 58 $b 2 $i 2000
No means is provided in subfield $y to express enumeration values that increment with a value greater than one. Since subfield $y specifies regularity in terms of chronology only, based on a chronology code expressed in the value of this subfield, it is not possible to define prediction on titles in terms of enumeration that skips the intervals between even or between odd numbers.
2.1 Define a new 853-855 subfield, Subfield $s
The definition of an 85x subfield (i.e., $s) that would function at each level of enumeration and permit the definition of an "incrementing interval" that would enable a process of identification and prediction for this kind of sequencing activity should be considered. Many serials with alternating sequential numbering occur in the scientific and technological fields. For such titles, the new subfield $s, can provide a vehicle to anticipate different conventions in the numbering patterns of serial issues. This new subfield would also serve titles described exclusively in terms of their enumeration and would make allowances for those whose regular range of volume numbering calls for intervals of two or more in calculating the next predicted issue. An example is provided in the "International Journal of Cancer". This title has three basic bibliographic units: "Journal international du cancer", "Predictive Oncology", and "Radiation oncology investigations".
Journal international du cancer was issued from v.85-88 in year, 2000,
Predictive Oncology issued v. 89 in year 2000,
Radiation oncology investigations issued v.90 in year 2000.
Presumably, in year 2001, the following sequential issues can be expected:
Journal international du cancer was issued from v.91-94 in year, 2001,
Predictive Oncology issued v. 95 in year 2001,
Radiation oncology investigations issued v.96 in year 2001.
If we take the simplest of these components, Predictive Oncology, we could foresee a predictive 853 field, using Subfield $s, that could anticipate its behavior:
853 02 $a v. $s 6 $b no. $u 6 $v
r $i (year) $j (month) $k (day) $w b
863 02 $a 95 $b 2 $i 2001 $j Mar. $k 15
Subfield $s is the number you add to the current enumeration value to predict the next issue when that value is greater than one. This subfield recognizes instances where enumeration values skip. In this example my last issue of Predictive Oncology had a first level enumeration of "89" in the year 2000. With the value "6" in subfield $s added to 89, the prediction is able to forecast a volume number of "95" for the 2001 volume year on Predictive Oncology. This method of coding allows for titles whose numbering continuity ($v) is continuous as well as for those whose numbering continuity re-starts.
2.2 Enhance Subfield $y of the 853-855 fields within each enumeration level to include a new code for enumeration regularity.
It might be considered whether subfield $y could be adapted to allow for expressing regularity in terms of enumeration rather than only through chronology. Subfield $y includes codes for the following:
1st position: Publication code (o=omitted or p=published)
2nd position: Chronology code definition (d=day; m=month, etc.)
3rd position: Chronology type code (e.g. d=dd; m=MM; d=MMDD)
These are used to interpret the regularity codes that follow, which are variable length.
If the title offered earlier is reconsidered:
Section A: v.1, no. 1, v.1, no. 3, v.1, no. 5, v.1, no. 7, v.1, no. 9, v.1, no. 11
To adapt subfield $y of the 853-855 fields to express regularity in terms of enumeration rather than chronology would involve defining new codes within subfield $y to designate that the rest of the field expresses regularity in terms of enumeration rather than chronology. To show that the enumeration skips could be accomplished by encoding the explicit enumeration that is provided for in this regularity pattern.
Codes would need to be assigned that identify which level of enumeration is shown by the regularity pattern as well as whether it is primary or alternative enumeration. It would be important to use fixed length codes for elements corresponding to chronology code definition and chronology type code, since the contents of the subfield would need to be interpreted in the same way that the existing subfield $y chronology codes are. The following could be defined:
1st position: Publication code (presumably always p, since the enumeration in terms of what is
published is given explicitly)
2nd position: Enumeration code definition (e=enumeration, primary; f=enumeration, alternative)
3rd position: Level of enumeration regularity (e.g. 1, 2, 3)
853 02 $a v. $b no. $u 6 $v r $i (year) $j (month) $k (day) $w b $ype2no.2,no.4,no.6,no.8, no.10.
This method could only be applied in instances where the values in subfield $y were characterized by a numbering continuity ($v) that re-starts and does not cover the common instances where enumeration is continuous. Thus, for titles like Lancet whose second enumeration levels do not re-start it would not be possible to encode enumeration-only regularity since the numbering cannot be reliably predicted. This is due to the fact that the enumeration does not assume the form of a repeating pattern but has different values every year. For titles with frequencies like weekly or bi-weekly, the string of subfield $y data could be very long in cases that cover skipped numbers. For those enumerations that are covered, it might be possible to code both for the enumeration regularity and the chronology regularity by using a repeated subfield $y. If we change our earlier example to a five times per year title, we could include:
Section A: v.1, no. 1 (2000, Jan.), v.1, no. 3 (2000, Mar.), v.1, no. 5 (2000 May), v.1, no. 7 (2000 Jul), v.1, no. 9 (2000 Sep/Nov)
We would have the option of expressing the regularity ($y) in terms of both its enumeration and chronology by repeating the subfield.
853 02 $a v. $b no. $u 6 $v r $i (year) $j (month) $k (day) $w 5 $ype2 no.1, no.3, no.5, no.7, no.09 $ypm 01,03,05,07,09/11
This example illustrates a case where the chronology combines September and November at the second level and relates that combination to number 9 in the enumeration.
3.1. The provision of an incrementing interval for enumeration allows for effective prediction of intermeshed enumeration, but assumes that chronological elements do not require a similar level of combination and synchrony with enumeration. Can the introduction of a new subfield within levels of enumeration provide relief for predicting such titles? What complications are likely to occur in meshing multiple layers of intersequenced issues across different components of a title?
3.2. If a new subfield is defined, is there a need for subfield $s at all levels of enumeration?
3.3. Should a change to subfield $y be considered to accommodate "skipped" issue numbers? What sorts of code definitions could be used?