FINAL REPORT: BIBCO Working Group on Series Numbering

(Approved by the PoCo (with guidance for execution of the recommendations), November 1999)

Introduction -- Results of statistical tabulations from the survey -- Summary of the important issues raised in the narrative answers -- Recommendations

Background

AACR2R rule 21.30L1 states, "Make an added entry under the heading for a series for each separately catalogued work in the series if it provides a useful collocation. Optionally, add the numeric or other designation of each work in the series." The corresponding LC rule interpretation, under 21.30L, states "Option decision. Apply the optional provision of 21.30L1 for adding the numeric, etc., designation of the series, if present in the item, in the series added entry. Give it in the form established on the series authority record."

Current PCC procedures require members to follow the national-level decision for form of numbering which is given in the series authority record (SAR) field 642 as $5 DPCC. BIBCO and NACO members are also permitted to record their local decisions in the SAR. If a BIBCO library chooses not to follow the national-level decision, the library has the option of not designating the bibliographic record as PCC.

The National Library of Medicine proposed that BIBCO libraries should not be required to use the 490 1/8XX combination when the only difference in the series is in the form of numbering. NLM proposed that when the only difference between the authorized form of the series and the form of the series statement on the item-in-hand is the form of numbering, one should not have to go through the 4901/8xx formulation. In the interest of efficiency, NLM asked that the PCC reconsider this requirement.

For a more complete statement of the situation, see Appendix 1

Introduction

The Working Group on Series Numbering (WGS) was appointed in late August 1999 to make recommendations by October 1999 on the series numbering issue described above. Members of the group are: Alice Jacobs (National Library of Medicine), Simone Jones (Library of Congress), May Rathbone (University of Washington), John Riemer (University of Georgia), Lois McCune Sewell (Indiana University), and Andrea Stamm, chair (Northwestern University).

The WGS communicated by electronic mail. In order to reach the broadest possible audience in a short time, the WGS designed a survey Appendix 2 which was sent out to the PCC list and to AUTOCAT. BIBCO liaisons were also individually contacted and requested to complete the survey.

• The WGS received a total of 71 responses, 69 of which completed the actual survey.
• Of the 37 BIBCO libraries, 34 replies were received.

• 52 university/college libraries (29 BIBCO)
• 7 public libraries (3 BIBCO)
• 5 government libraries (4 BIBCO)
• 6 special libraries not affiliated with universities (2 BIBCO) and
• 1 library organization.
• 3 of the respondents were international libraries.

Results of statistical tabulations from the survey

Important issues:

1. The feasibility of discontinuing the practice of explicitly tracing a series in an 8XX field when the only difference between the transcription of the series statement and the added entry is the form of the series numbering.
   We asked respondents which alternative they would favor (question 7):
   
   • 14 (20.3%) favor dropping the 490 field and using only an 830 field with the series authority record (SAR) 642 field authorized volume number and caption/designation;
   • 24 (34.8%) favor using the authorized form of numbering (from the SAR 642) in the 440 regardless of what is on the item;
   • 9 (13%) favor using the volume number and the caption/designation of the item-in-hand in the 440, regardless of what is on the SAR 642 field;
   • 14 (20.3%) favor dropping the caption/designation entirely and using only the volume number or letter;
   • 8 people (11.6%) favor none of the alternatives.

   Numbers alone do not depict the whole picture here. Many respondents have written comments in other part of the survey that clearly indicated that they considered all the suggested changes to be problematic. Many people chose one of the choices but appeared to be unenthusiastic about their choice. Looking back, the WGS should have given 'none of the above' as another alternative in question 7.

   The factors which influenced respondents' choices to question 7 yielded interesting results (question 8). Many respondents chose three to four of the possible choices. 42 (60.9%) chose cataloging efficiency, 51 (73.9%) chose impact on the user, and 41 (59.4%) chose impact on public services staff. Less popular choices were the need for an accurate record of what appeared on the piece, chosen by 16 (23.2%) people as well as "other" chosen by 18 (26.1%) people.

2. The impact that the discontinuance of this practice might have on OPAC displays.
   The WGS asked the respondents how their local systems sort a series display (question 2). 34 (49%) respondents have a library system that sort series titles alphabetically, then by series subfield $v. This implies that if the PCC adopts alternative #7c, and possibly #7d, if a library does not change all its retrospective bibliographic records, the series will not sort in the proper display order.

   The WGS asked if an inconsistent use of the caption/designation in series subfield $v causes a problem for the public service librarians and gave an example of such inconsistent use in question 3. 46 (66.7%) of the respondents said that the inconsistent display causes a problem. Among these 46 libraries, 28 (40.2%) consider this problem as unacceptable or it matters a lot to their users. Only 9 (13%) respondents consider this problem mattering a little. The rest either gave no answer, or provided various comments from "no idea" to "sorting matters a great deal however, with series of which we own a large number of titles as it becomes far more difficult to find a particular number of series." (Tulane)

   We also asked how many library systems are able to sort series on volume number rather than on the caption/designation used in the subfield $v (question 6). 59 (85.5%) said that their systems do not have such capability, while 9 said "don't know." Thus current programming across nearly all library systems does not ignore the caption/designation in subfield $v so the series titles do not display in the correct order. For a few libraries, the library has already implemented the equivalent of alternative #7d, where the cataloger currently strips out all captions/designations in the series subfield $v to achieve the correct sort order.

   The WGS also asked the respondents what consequences there might be on local systems if the PCC adopts one of the alternatives in question 7 (question 9). 14 (20.3%) respondents indicated that there would be consequences if any one of the alternatives were adopted. 18 (26.1%) indicated that there would be no consequences at all. From the responses, the alternative which would have the most impact on library systems is #7d (36 respondents (52.2%)), while #7b (20 respondents (29%)) would have the least impact on library systems.

3. The impact that the discontinuance of this practice might have on timeliness of cataloging.
   In order to evaluate the impact of the alternatives on the timeliness of cataloging, the WGS asked respondents to evaluate each alternative regarding its impact on the timeliness of cataloging (question 10). 29 respondents (42%) said that none of the alternatives would affect the timeliness of cataloging in their libraries while 9 respondents (13%) said the opposite, that any one of the alternatives would affect the timeliness of cataloging in their libraries. Among the 69 responses (each with 4 possible answers), there were 74 "yes", 156 "no", and 46 with no clear answer.
   

   We interpret the 156 "no" answers to mean that most of the time, people do not think any of the alternatives would cause any impact on the timeliness of cataloging.

   For the majority of the "yes" answers, respondents said that the proposals would negatively affect the timeliness of cataloging and also would increase maintenance work. There were very few people who answered "yes", with the meaning that it would positively affect timeliness of cataloging.

4. Impact on the BIBCO program.
   We asked all respondents if the PCC adopts one of the changes, should it be optional (question 11)? 37 (53.6%) said it should be optional, while 28 (40.6%) said it should not be. 34 of the 37 BIBCO libraries listed on the PCC liaison web page responded to our survey. Out of the 34, 23 (67.6%) said that if the proposed change is not optional (i.e., it is required), it would not affect the amount of PCC core records that they would contribute. Most respondents did not give a reason, but for some of those who did, the reason is that they do not contribute any core records. One additional respondent, mentioning the inadequacies of core records, is also likely not a PCC core contributor. 9 libraries said that the change would affect the amount of PCC core records they would contribute-- 7 libraries (20.6%) said by a small amount and 2 libraries (5.9%) said by a significant amount.

   The WGS asked if current BIBCO practice is not followed (i.e., the series subfield $v does not follow the form of series designation in the national SAR 642), how much is the overall quality of a bibliographic record adversely affected (question 15)? 17 (24.6 %) said that the quality is unacceptable, and 17 (24.6%) said the quality is affected a lot.

   Adding those two answers together comes to 34 (47.8 %) having a strong negative reaction to the proposals as they impact on the quality of bibliographic records. 32 (46.4%) respondents said the quality is affected "a little" or "not at all." We noticed that 19 out of the 32 respondents' local systems do not sort series by the subfield $v. This tells us that most people were thinking of their local systems' current capabilities when they answered the survey questions. They were not thinking of the effects this proposed change might have on the library community and its users.

5. The concept of series transcription vs. access points.
   Based on the responses for question 7, 87% of the respondents chose alternatives #a, b and d, while only 13% chose #c. Alternative #c is the only alternative that allows the exact transcription of the series statement. This result tells us that if libraries have to make a choice, they would choose to have the correct access points over a correct transcription. However, several respondents were adamant about the need to retain the actual transcription of the series numbering in the item-at-hand (see point 7 in the section of narrative comments).

6. Other issues.
   The WGS asked if the respondents think that library system vendors should help remedy the problem with series sorting (question 13). There was a clear mandate here: 62 (89.9%) answered yes to the question. Unfortunately, the survey question did not ask further if vendors should be asked to sort the subfield $v on numerals exclusive of caption(s) used or sort on captions that (hopefully) have been consistently entered.

   The WGS asked the respondents if they think that MARC21 fields 440, 490, 8XX, should be changed in some way to delineate a mechanism to improve series sorting in library systems (question 14). While 31 (44.9%) respondents answered yes to that question, 20 (30%) said no, and this latter group gave very convincing reasons why a new subfield for volume designation should not be defined. 15 other respondents (21.7%) gave various comments rather than a straight "yes" or "no" answer. Their comments reflect a lot of doubt on what a new subfield could accomplish, and express concerns about the update and maintenance that they need to do to the retrospective bibliographic records.

Summary of the important issues raised in the narrative answers

General statement: few respondents showed more than lukewarm enthusiasm for the proposal, and many were adamantly opposed to it for a variety of reasons. These reasons include:

• loss of proper transcription data;
• problems in OPAC displays, especially when there are many bibliographic records associated with the series;
• problems with users, ILL, acquisitions, and reference staff in missing the proper citation due to illogical sorting;
• loss of trust with PCC and the need for copy catalogers to check what is now thought to be an LC-quality record;
• no significant time saving, especially when most systems have copy-and-paste capability.

1. Problems in OPAC displays would result from adopting each of the proposals, and the problems seemed to depend partially upon what types of library systems the respondents use. There were four choices given in question 7, and here are some of the problems that various respondents associated with each choice.
   1. Drop the 490 and use only an 830 field with the series authority record (SAR) 642 field authorized volume number and caption/designation.
      • Some systems won't allow an 830 without a 490 field
      • One would lose accurate transcription of the subfield $v information, and sometimes that is very important in solving problems, especially on different series which have the same name or very similar names.
      • With copy and paste capabilities so common in many systems, there is no need to change the rule just to shave a few seconds off creating the record.
   2. Use the authorized form of numbering (from the SAR 642) in the 440 regardless of what is on the item.
      • Respondents pointed out that this alternative was very similar to choice A, and that the ramifications are the same: loss of data of what really was on the piece, more difficulty in solving problems, and little actual time saved when copy and paste can be used.
   3. Use the volume number and the caption/designation of the item-in-hand in the 440, regardless of what is on the SAR 642 field.
      • This was the least popular choice of respondents; only 9 made this choice. In systems which index by publication date, it was not as troublesome as in those which sorted by title within the series or by volume number (where it was felt to be unacceptable by many).
      • Respondents felt that this was the most disastrous and trouble-causing of all the choices. Some said that sorting in their catalogs would be chaotic, and others cited strange looking displays in their OPACS.
      • For series where the number comes first and then the date, it would be very confusing to not be able to reverse these elements. For example, 1/1989 becomes 1989/1 in the present rules. Without the ability to reverse these, the sort order could become 1/1989, 1/1990, 2/1989, 2/1990, etc.
   4. Drop the caption/designation entirely and use only the volume number or letter
      • Although 14 respondents made this choice, they may not have been aware of the ramifications of it. Respondents from large (generally academic) libraries pointed out that for complex types of numbering, this would cause confusing or meaningless displays. For example, a German series might have Reihe, Abteilung, Bd., and further subdivisions. If these are dropped the display could look like this: 1, 3, 2, 5 which is very confusing.
      • Even if this were adopted, there are millions of records in databases where the current numbering still exists, and this would create sorting problems and database cleanup problems. Libraries do not have resources to clean these things up.

   Note: even though the survey failed to give this as an option, seven libraries replied to question 7 that none of the options was acceptable. Had this been an option, the WGS believes that even more libraries would have chosen "none" as their answer to this question. These libraries also stated that if any of the proposals was adopted, they would edit the copy to conform to our current practice.

2. The change would adversely affect the copy cataloging workflow. Many respondents noted that the PCC records are considered to be just like LC records in their workflows. If the proposal were adopted, they would probably have to move these from the "fast cataloging" track to the catalogers who deal with more complex copy. BYU mentioned that adopting any of these options would slow down cataloging because the theory of BIBCO records is that you can find reliable and usable records in the database. With the change, the records would require checking and possibly editing. This numbering issue arises in only a small number of cases, and it is not that hard to deal with in a 490/8XX combination. But with the change, all records might be suspect. Thus the timeliness of cataloging would be affected.
3. Many of the respondents mentioned that the integrity of PCC records would be compromised. Someone wondered why the 4XX field should be less important than any other field transcribed in the record. The proposal breeds mistrust of copy cataloging and invites scrutiny of copy.
4. Many felt that no time would be saved, and that in fact, the time saved on the front end would be wasted many times over on the other end.
5. Many cited any of the choices as being a burden on database maintenance and cleanup. Some believed it would be like the subfield $v in subject fields -- we would just go forward from here, and not bother to fix differing forms in databases. This would result in inconsistencies.
6. There is the issue of inconsistency if only the BIBCO members adopted one of the proposals, and the rest of the cataloging community did not.
7. There is the issue of the dual nature of series headings (transcription vs. access point). BYU stated, "The current practice of creating 490/830 even for numbering changes is clear and easily applied. It works well. We can transcribe the series headings exactly as they appear and if we chose to index them the way they appear or index them differently we can do that as well." Gary Strawn of Northwestern proposed using a 490/8XX in every instance to address this transcription/indexing issue.
8. There seemed to be consensus on question 13, "Do you think library system vendors should be asked to remedy the problem with series sorting in online public access catalogs?"
   Most replied "yes," although one said, "I really think they have better things to do." Some believed that their vendors were responsive to their requests, and others did not. Gary Strawn of Northwestern said, "Yes, obviously. It is the system vendors who have created the problem through their arrogant disregard for obvious needs, or through their failure to take any action at all. There are automated solutions available that would admirably suit 90% of cases." A related thought from Oberlin was that this was the tail wagging the dog. "The limitations of this or that ILS vendor should not dictate cataloging policy or practice."
9. Regarding question 14, concerning a way to delineate a mechanism to improve series sorting in library systems (for example, a specific subfield for volume designation in series which might be defined to assist library vendors in providing support for series sorting) the responses were varied.
   Some thought it was a good idea, but others gave reasons why it was not. Queens Public Library said, "There already is a volume subfield in MARC and adding another subfield of some sort would involve a great deal of database maintenance to deal with all the series already in the library." Gary Strawn of Northwestern replied, "Not until a system solution has been explored first. Whatever solution to the problem is devised must be capable of handling millions of existing records without change." From Oklahoma State Library: "It complicates tagging. I would like to see that library vendors develop systems to skip the volume designation in sorting series." And from John Riemer of the University of Georgia: "If it bears research, research it; however, do not change without a known benefit to be derived."
10. If any of the options to question 7 were adopted, there would be an impact on the BIBCO program from two avenues.
    • First, many libraries, especially BIBCO libraries, would edit copy, thus affecting the timeliness of cataloging.
    • Second, two of the respondents stated that any of the proposals would also affect, by a significant amount, the number of core records contributed to the BIBCO program while seven BIBCO respondents stated that it would affect their core contributions by a small amount. Two respondents stated that they did not contribute core records, so they did not answer this question.

Recommendations

1. No change should be made in the existing BIBCO practice of explicitly tracing a series in an 8XX field when the only difference between the transcription of the series statement and the added entry is the form of series numbering.
2. LC should issue an LCRI to AACR2R Appendix B.9 which adds "v." as the abbreviation for "vol.". The effect of this RI would be to reduce the number of times a 490/8XX combination would be needed in cataloging.
3. The PCC should contact vendors to work on correcting all numerical sorting of series entries in the OPAC. Support for series sorting (through the entire numbering) is desirable in an integrated library system. Methods already suggested include asking for the input of each vendor's user group as well as having the PCC write a letter to each of the major vendors.
4. MARBI and vendors should work towards developing a mechanism which supports disregarding the designation in the series $v in its sort of the series.

The 490/8XX proposal: The WGS could not agree to include this last issue as a recommendation but felt that it should still be mentioned in our report. Several members felt that there was merit in the proposal while at least one member felt that the proposal is contrary to the NLM proposal of cost-effective cataloging.

The PCC and MARBI should consider investigating the possibility of recording ALL series with a 490/1 and an 8XX. This proposal would clearly separate out the needs of series transcription from access points which are currently mixed together in the 4XX fields. We understand that this proposal was put forward some years ago by Mary Dabney Wilson.

With the advent of copy and paste capabilities in most online catalogs today, there is new incentive to clear up the transcription vs. access issue. Because of the possible burden placed on existing staff, the clean up of retrospective records is not essential and leaving them 'as is' would cause no harm. This is the same situation as we are now faced with the subfield $v in subject headings. The advantage of this proposal would occur in the future when files of series were in need of maintenance. It would be easier to change the 8XX only, rather than our current practice, which is if the 8XX form exactly matches the 490, we delete the 8XX, retag the 490 to a 440, and insert the appropriate subfield codes such as the $x and $v, etc. Anyone who has experienced the manual clean up a large file of series added entries at the time of AACR2 implementation has some understanding of the time savings of using automation to clean up 8XX added entries only.