J. McRee Elrod
Director
Special Libraries Cataloguing, Inc.
4493 Lindholm Road
Victoria, B.C., Canada
V9C-3Y1

Summary

During most of this century, cataloguers have constructed catalogues using a generally accepted configuration of data as exemplified by the unit card. This configuration was later codified as the International Standard Bibliographic Description (ISBD). Skill in using a catalogue was generally transferable from one card catalogue to another. With the creation of on-line public access catalogues (OPACs), catalogue construction passed from the hands of cataloguers to systems designers. Commonality of search techniques, and consistent display of bibliographic data, vanished. Cataloguers should resume their historic role in catalogue construction, and restore a minimum display of core bibliographic data to library catalogues.

There was a time when we as cataloguers were concerned with building catalogues. Now we seem primarily to be concerned with creating bibliographic records, leaving catalogue construction to others.

At a recent library conference, I sat in on a committee with the mandate to propose minimum standards for a particular type of material in a particular type of library. No member of the committee was willing to consider either requesting to have the mandate extended to consideration of the display of that standard record, or to consider a further mandate for another committee to consider questions of display. As one cataloguer put it, the construction of the catalogue was out of her hands. Such blinkering of concern seems quite sad to me.

It is not enough for us to create good bibliographic records. We should see that search mechanisms provide an intelligible mode of access to those records we have created, and that those records are displayed in an intelligible manner.

It is time for cataloguers to return to determining how the records we create are seen by the public. We cataloguers have a standard for the bibliographic record, the core of which has been remarkably stable for the past century. A 1901 Library of Congress printed card is as readily comprehensible today as then, and its central core does not differ greatly in the information given, and the order in which it is given, from the central core of a 1995 MARC record. We have a long tradition of catalogue building upon which to draw. We now have the powerful new tool of being able to search the full text of a bibliographic description using key or any word techniques, as well as searching by the traditional access points; and we have the ability to show information not before available, such as circulation status.

The primary problem is that many OPAC displays deconstruct the bibliographic record. One example of this is the practice of some OPACs of displaying 245$a$b (title proper and subtitle), followed by all 1XX and 7XX (main and added entries) labeled "authors". Just what these "authors'" relation to the work might be can not be determined since 245$c (statement of responsibility) is nowhere to be seen.

We as cataloguers create a coherent record which makes apparent the nature of the work being described, only to have it fragmented by software creators who have no concept of cataloguing tradition, or the needs of library users, and see the record as a collection of modular bits to be shifted about at will.

Gregory J. Wool (1993) and others in comparing bibliographic descriptions composed using ISBD with their OPAC version displays, found that "changes ... involve rearrangement, inaccurate labeling, repetition, addition, and omission of data elements along with the elimination of the traditional distinction between descriptive data elements and accesses points." The authors suggest that current descriptive cataloguing standards may be ill-suited to the creation of records for online display, and suggest new rules. If recommendation is carried to its logical conclusion, we would be creating new rules to facilitate current ill conceived display designs. I would suggest, rather, that current OPAC practice is making a hash of perfectly adequate bibliographic descriptions.

In contrast to the cataloguer mentioned at the beginning of this paper, who felt that bibliographic display was "out of her hands", one librarian at the same conference was found (in private conversation) who was actively engaged in catalogue construction, and was experimenting with different displays and catalogue configurations. Michael J. McGuire of the Law Society of Saskatchewan has made available to his patrons a classed* browsable file, i.e., a file in call number order, which may be scanned in any area of the classification. His users find it quite helpful. This might be an option to be considered by libraries as more material goes into off site storage, and shelf browsing gives access to a decreasing portion of the collection.

Another simple way to improve user access to bibliographic information is to subsort the subject index by date.

Mr. McGuire accepted my suggestion that (using the "date one" fixed field) retreived items for subject searches be in inverse chronological order, making a more useful display than his present alphabetically arranged one. He recognized that an inverse date order would call the most recent imprints to users' attention, particularly in large files in law collections such as "Environmental Law". He has mastered his OPAC software, and can make such changes himself without recourse to vendors or computer technicians. (Having a PC based system, he has access to the source code.) For those libraries dealing with turn key automated systems, concerted cooperative effort must be brought to bear on system vendors to make such changes.

Leaving the issues of search mechanisms and help screens which have been better served by published research (Campbell 1992, O'Brien 1994) than bibliogrpahic display, let us consider OPAC displays in relation to our long standing tradition of making comprehensible bibliographic records available to users. Some published research has addressed this question (Shires 1992, Crawford 1987 & 1992, Matthews 1987). Perhaps most noteworthy is that Gregory J. Wool and others (1993), reporting a study conducted at the University of Nebraska- Lincoln Library, comparing cards and online versions of records. Walt Crawford's (1987) report of a study conducted for RLG proposed bibliographic display screens which are understandable, attractive, and compact.

Juliana Chan (in a 1995 reearch project for the MIS degree at the University of Toronto) found, applying a check list developed from the published literature on OPAC displays, that the average score for OPACs in selected academic libraries in Canada was only 59%, and for public libraries was only 55%. Even the research which has been done is not being applied. Chan wonders if this wide gap between displays existing in OPAC systems, and the ideal display as developed by exisiting research, may be be the result of programmers not being aware of the published guidelines for displays, and not having sufficient expertise in the library field to fully understand library terms. It would seem to be the task of cataloguers to impart this information.

The heart of the bibliographic description is that main body of the old unit card, title through collation, which is also the heart of the later standard, the ISBD. We should not lightly break it up. I would suggest that even a one line description in an OPAC can usually display much of 245$a$b$c (title, subtitle, and statement of responsibility) and 260$c (publication date), if space is not taken up by labels. (Any labeling of brief displays could be done at the head of columns.)

When moving to a fuller level of description, 245 through at least 300 or 440 (title through collation or series) would seem to be a coherent body of information. (Of course for older records created under ALA rules in which statement of responsibility was not given if it is the same as the main entry, the main entry would need to be displayed in lieu of 245$c.)

Labeled access points, I maintain, should not replace descriptive data. A 245$c "edited by", "collected by", "translated by", "in honor of", "sponsored by", etc. is far more helpful than having these persons' and corporate bodies' access points listed as "authors". (Keeping description distinct in our minds from access points is becoming more difficult since their functions are increasingly being combined in single MARC fields such as 246, 440, 780, and 785.)

Walt Crawford (1992) is eloquent about those things an online catalogue can do, that a card catalogue can't do, and one of those is labeling. But even he admits that it is difficult to come up with labels for some fields which are always appropriate. He also admits, as his RLG study (Crawford 1987) found, that the introduction of labels will usually increase a one screen display to two for most full records. His basic principles for how much information should be on one screen, and the use of space, are very helpful. These principles have been given clear presentation by Nancy Lee Shires and Lydia P, Olszak (Shires 1992). One important principle is that screens should reserve standard locations for certain information, and be consistent. (Crawford, Shires, and Olszak all neglect to mention how well a unit card met that and other principles they outline!)

Although rules for bibliographic description may be quite adequate, now may be the time to consider enhancing cataloguing rules to address some of the questions which have arisen concerning the display of data in OPACs. The focus of this enhancement should be to indicate that the ISBD core be considered a unit for all but the briefest displays, and not lightly to be fragmented. Perhaps rules, or a separate display standard, should mandate some of the design principles Crawford outlines. Crawford (1987) further suggests the labels "Title" (245$a$b$c), "Published" (260$a$b$c), and "Material" (300$a$b$c) for the main body of the description. In "Testing Bibliographic Displays for Online Catalogs", one of the very few studies to actually test patrons' response to a variety of bibliographic display screens, he found that right justified labels with left justified data was the most easily understood. His suggested screens (Crawford 1987 p. 28) meet his stated objectives of understandability, attractiveness, and compactness.

I do not find labels helpful for the core ISBD display. I would suggest the following labels for other elements of fuller displays:

100, 700, (and 600?)      Persons

110, 710, (and 610?)      Corporate bodies

111, 711, (and 611?)      Conferences

The remaining labels are less problematic:

4XX, 830                  Series

5XX                       Notes

6XX                       Subjects

To return to how we construct our catalogues, an earlier experience demonstrated to me that most of us tend to prefer what we know in catalogue design, as in most things. During the process of dividing a dictionary catalogue in an academic library, we experimented with filing personal name subject headings as persons in the undergraduate library author/title catalogue, and as subjects in the main library subject catalogue. The idea was that with a year's experience, we would select one pattern for the entire campus. To my astonishment, neither library wished to give up its own arrangement. Since we do tend to like what we know, it is difficult to achieve unbiased results in studies of catalogue configuration.

Each library is of course free to construct the catalogue which best serves its own patrons. But we must begin comparing experiences and try to arrive at some consensus and commonality among catalogues. We should have an agreed upon core of search techniques for locating material by the traditional access points, and a standard practice of displaying the ISBD description. Enrichments and experimentation should be additions to, not departures from, that standard.

Joseph Blackburn (hlcat@ttacs.ttu.edu)) in a posting on Autocat (an Internet discussion list) stated that he considers one of the most important issues facing our profession is our influence over the display of bibliographic records on automated systems. He feels we only have the illusion of controlling how the records we create will be displayed. He continues, "I have been to several format integration workshops. At each, the general response was 'Oh, that's very nice, but we'll only be able to do what out SYSTEM allows us to do'. ... Wouldn't it be nice if we had a profession-wide set of minimum standards for OPACs, e.g., 'must be able to display notes in any specified order, regardless of tag number'? Couldn't we then use our pocket books to require systems operators to comply?"

Conclusion

In conclusion, our cataloguing rules may have to be modified to address the question of optimal display of information. The ability of systems to deconstruct the bibliographic record we have constructed is relatively new. The coherency of the unit card is little more than a fond memory. We as cataloguers should be able to return that coherency to bibliographic displays in OPACs. Search mechanisms and help screens might be more subject to experimentation, and need more influence from our reference librarian colleagues, but some consensus is needed there as well, just as we reached consensus concerning the provision of cross references in card catalogues.

Not all catalogues need be identical, just as card catalogues were not identical. We had dictionary catalogues, divided catalogues, and ticked tracing guide carded catalogues. There was, however, a commonality in the way information was presented which allowed patrons to move easily from one to another. OPACs in their bewildering variety have existed long enough for us to begin to impose some order upon them, first in terms of bibliographic displays, and then in basic search methods, based on our long experience of presenting bibliographic information to library users. These understandings should be incorporated in a standard such as Mr. Blackburn proposes, and we should purchase only those OPAC software systems which adhere to that standard. For too long we have passively accepted what is handed to us, which has often been created by engineers and system designers based on what they perceive that library users will want, rather than the well established and known needs of users to locate information, which we have learned over generations.

Robert Fulghum tells us he learned all he really needed to know for life in kindergarten. While kindergarten might have been a bit early, by the end of elementary school I had learned the card catalogue skills which served me through graduate school, and in public and corporate libraries thereafter. Now I must learn new techniques in any library I enter. The search skills I develop in a nearby university library are no help in my local public library. The directions on my local bank cash machine are clearer, and have more in common with those on other cash machines, than the various introductory screens of the OPACs I use. The configuration of the bibliographic screens are equally varied. Nicholson Baker (1994) expressed very well the response of one scholar to the loss of this carry over of experience from one library to another. We need to return to having catalogue use skills fully transferable among libraries. This would be best accomplished, I am convinced, if we cataloguers returned to catalogue building.

Note

(*) Mr. McGuire has the advantage of Moys (A classification scheme for law books / by Elizabeth Moys. -- London : Butterworths, 1968) for his classed display. Moys classification does for law what NLM (National Library of Medicine classification) does for medicine, create a broad subject arrangement which reflects the use of material by professionals in the field. Moys also arranges common law subjects together by subject, rather than first by jurisdiction. Moys uses the letter K, and like NLM's W, may be shelved with books in the LC classification. Moys also has a replacement for DC's 340.

Works Cited

Baker, Nicholson. 1994. Annals of scholarship : discards. The New Yorker 70, no. 7: 64-86.

Basista, Thomas et al. 1991. Designing the OPAC user interface to improve access and retrieval. Microcomputers for information management 8, no. 2: 87-103.

Chan, Juliana. 1995. An evaluation of displays of bibliographic records in OPACs in Canadian adademic and public libraries. Toronto, Ont. : University of Toronto, Faculty of Information Studies, 1995.

Campbell, John et al. 1992. Online Catalog Documentation Task Force : Onscreen Documentation Subgroup final report. University of Georgia.

Carter, Trina and Park, Hye Ok. 1993. A user survey of the online public access catalog at California State University, Fresno. Library software 12, no. 1: 43-53.

Crawford, Walt. 1987. Testing bibliographic displays for online catalogues. Information technology and libraries 6, no. 1: 20-33.

Crawford, Walt. 1992. Starting over : current issues in online catalog user interface design. Information technology and libraries 11, no. 1: 62-76.

Dramenstott, Karen M. and Weller, Marjories S. 1994. Testing a new design for subject searching in online catalogs. Library hi tech 12, no. 1:67- 76, 86.

Fayen, Emily. 1986. You've come a long way baby, but ... . University of Pennsylvania.

Knutson, Gunnar. 1991. Subject enhancement : report on an experiment. College and research libraries 52, no. 1: 65-79.

Matthews, Joseph R. 1987. Suggested guidelines for screen layouts and design of online catalogs. Library trends 25, no. 4: 555-570.

O'Brien, Ann. 1994. Online catalogs : enhancements and developments. Annual review of information science and technology 29: 219-242.

Shaw, Debora. 1991. The human-computer interface for information retrieval. Annual review of information science and technology 26: 155- 195.

Shires, Nancy Lee and Olszak, Lydia P. 1992. What our screens should look like : an introduction to effective OPAC screens. RQ 31, no. 3: 357-369.

Wool, Gregory J. et al. 1993. Cataloguing standards and machine translation : a study of reformatted ISBD records in an online catalog. Information technology and libraries 12, no. 4: 383-402.