Committee on Cataloging: Description and Access

Task Force on the Review of the IFLA Guidelines for OPAC Displays

Final Report

Mitch Turitz, Chair
Matthew Beacom
Gabriele I. Kupitz
Martin Kurth
Shirley J. Lincicum
Aimee Quinn
Barbara Rapoport

Please note that the purpose of these documents is to facilitate the work of the Committee and to provide a means for outreach to both library and non-library cataloging communities. The documents are intended for the exclusive use of CC:DA and its cataloging constituencies, and are presented for discussion in the ongoing process of rule revision. Under no circumstances should the information here be copied or re-transmitted without prior consultation with the current Chair of CC:DA.

The Task Force on the Review of the IFLA Guidelines for OPAC Displays has reviewed the draft Guidelines comparing the principles and recommendations to the functions of the catalog as expressed in the Paris Principles (“Statement of principles adopted at the International Conference on Cataloguing Principles, Paris, October, 1961”) and the Joint Steering Committee for the Revision of AACR2’s guidelines.

While the Task Force as a whole admired the Guidelines, different members had differing points to make. The Chair has attempted to represent these differing opinions in this text.

Overall

The Paris Principles largely address how and when catalogers should provide access points to items in the catalog, while the Guidelines address a different issue: how access points and records should be displayed to OPAC users. Although this does not mean the Guidelines conflict with the Paris Principles, the Task Force feels the introductory remarks of the Guidelines need to include a brief discussion of the differences between the world of the Paris Principles and the OPACS of today.

General points
1. The Task Force recommends that an overview of the 1961-1999 catalog be added to the Gudielines. The most significant features of the evolution of the online catalog, and its differences and potentialities from the catalog envisaged in the Paris Principles, should be described in one of the introductory sections or in Principle 1. If it is addressed in an introductory section, that should be referred to in Principle 1.
2. The function of “Ascertaining” as identified in the Paris Principles, is no longer always the end result of catalog searches: now a patron can sometimes reach the work itself through the OPAC. The Guidelines take the WWW, the Internet, and electronic resources in general into account when dealing with some display issues, but in others do not adequately identify the role of the OPAC as a portal that carries users from catalog records directly to the resources those records describe. Because of this relatively recent development, at least with regard to electronic resources, the object of the catalog has changed.
3. The Task Force feels the need for an introductory paragraph on the definitions of “author” “book” and “edition” (and also other major terms selected by the IFLA TF), to explain the expanded concepts embodied in these words. Discussion of such definitions can be found in some of the early footnotes – but they need to be brought together, and dealt with more prominently, rather than being scattered. Usage of these terms as defined should be followed throughout the document. The definitions section should also refer readers to the glossary for additional definitions. We recommend that terms defined in the glossary be put in bold or otherwise identified in the body of the document. See the Task Force’s comments on Principle 1 below. Additionally there was some discussion among Task Force members of using the words “item” instead of “book” and “version” instead of “edition.”
4. User options. The Guidelines frequently mention options that might be allowed in addition to the defaults that are the main focus of the Guidelines. Optionality should be expressed in the Guidelines as a principle, so that end users (not only library staff, vendors, and installation consultants) should have some control over catalog displays and over the records they retrieve.
  Some options may best be limited to library staff, but many display options could usefully be given to end users.
  Recommendations: The Task Force recommends that two new principles (our first choice) or general recommendations (our second choice) be added to the Guidelines:
5. Numbering of recommendations. The Task Force proposes a single numbered sequence for the recommendations, in order to make them easier to refer to. The Task Force recommends a table in the Guidelines that would group principles and their related recommendations by topic. This would offer an alternate form of organization of the Guidelines for readers who want to follow a specific strain of logic from one principle and/or recommendation to another.
6. Bulleting. Suggestion: bulleting for emphasis. In the Introduction to the Guidelines, particularly in the Scope section, using bullets would help highlight the main points. Highlighting main points with text bullets may aid the IFLA Task Force in clarifying their descriptions of the domain and structure of the Guidelines.

Editorial Suggestions for Principles

General points
1. Have at least one example for every principle (i.e. Principles 4-6 have no examples)
2. Add running headers especially when an example runs to the next page (see page 7 to page 8). A header could state on page 8 that this is a continuation of Figure 1, example 1, Results of a name search on Tristram Coffin.
3. All figures should show the term(s) searched and how they were searched. Some figures include this information, but Figure 2 on p. 16, for example, does not.
4. Graphical examples. We are moving into an era of graphical catalog interfaces, yet this document contains no “graphical” examples. It could benefit from some.
Comments on individual principles

Introduction: p.3: under the heading “Scope:”

“the Guidelines are intended to apply to all types of catalogue,” should be “catalogues”

Introduction: Scope (p.3)

One Task Force member comments: The “primary audience” referred to in this document is never explicitly defined. Most of my comments have been focused on “public” or end-user display needs since the “scope” statement in the introduction seems to imply that this is the primary focus of the Guidelines. In many cases, I think that staff (especially catalogers/tech. services staff) have somewhat different display needs than end users do since they use the catalog in a much different way than end-users do. The bias toward end-user, “public” display in this document should be explicitly stated, or more principles/recommendations should be added to address the particular needs of other major catalog user groups.
a word left out (?) on pg. 4 under the heading “Organization”, 2nd paragraph, lines 5-6: “However, it is recognized that under various circumstances it might [be] useful to group several . . . ”

Principle 1. (p.6)

First, we have to move beyond just “books.” The Guidelines leave out the footnote in the Paris Principles that explain that by the term “book” they mean just about anything one would find in a library collection. That is essential and should not be left out of the quotation. This is partly a PR thing, but saying that catalogs are for finding books alienates a number of communities among librarians and library users. (See in the Overall introductory section A.1.c.)
Recommendation: that the footnote defining “book” be added or the term should be added to the glossary. N.B.: the footnote suggested by the Task Force is the footnote that appears in the Paris Principles themselves (p. xiii of Statement of Principles . . . ). Without it, the meaning of the Paris Principles is altered.

Footnote 7 (p.14)

FRBR is referred to in the footnote but not defined either in the footnote or in the glossary

Principle 3 (p.14)

This principle should refer to the principle we have proposed on limiting retrieval sets. The statement that “All OPACs should be capable of summarizing, sorting and displaying thousands of headings and records quickly and efficiently” should be clarified by examples.

Principle 5 (p.15)

Recommendation: The title and scope of this principle need to be rethought because OPACs offer keyword-within-record searches, as described in the Organization section of the Guidelines. The title of the principle could be expanded to “Emphasize Author, Work, Subject or Other Search Terms Sought in Resultant Display” or simplified to “Emphasize Search Terms in Resultant Displays” The text of the principle as written suffices for searches that are clearly author, work, or subject searches, but fails to address keyword-within-record searches, which often do not clearly fall into one of these three categories. For example, a user entering a keyword-within-record search for the terms “information visualization” may intend to retrieve a specific work with those terms in the title, but may also be interested in retrieving records for other yet-unknown works that contain those terms in subject-related fields.
Reason: Display of a “keyword-within-record” search (4th searching mode, second type of search “keywords which the system matches without regard to order against all words in a single bib record” Such searches are not always on a subject, author or title access point, but on information in notes or other descriptive fields. Some searches could be primarily numerical.
Additionally: There is a potential conflict between this principle and Principle 25, which calls for ISBD order for single-record displays. The Guidelines should clarify that this principle applies to multi-record displays.

Principle 6 (p.15)

We would not want anyone to get the impression from this that they are not supposed to index subfield t in author-title entries at all, which is what some Task Force members thought the principle was suggesting until they read it more carefully.
Suggestion: This principle could benefit from more detailed examples.

Principle 8 (p.16) Punctuation.

Principles 8 & 9 state sorting elements should be respected. Should they also state that punctuation (e.g., parentheses and commas in subject headings) be respected in sorting, though not necessarily in searching?

Principle 9 (p. 19)

Principle 9 was the subject of a lengthy debate among Task Force members regarding “rotated headings.” There are some cases where automatically “rotated” headings are quite useful for patrons. Example: patrons searching for information on a particular geographic area. Rearranging subject heading strings does not always change the meaning of the strings. Example: Forests and forestry--Oregon means the same thing as Oregon--Forests and forestry, but only the former is now correct under LCSH. A patron could reasonably approach the topic from either a geographic or a topical point of view. In this case, although the DISPLAY of the rotated heading is not in the same order the cataloger intended, the catalog does not lose the original order of the subfields and display of the original heading. What happens is the system generates an additonal headings display (or a cross-reference without the cross). This is a recent technological innovation which was not possible at the time of the Paris Principles but needs to be differentiated from the concept of incorrectly reordering subfields.

Principle 11 (p. 20)

Some of the Task Force members agree with this concept of rotated headings only as far as this is a local or patron option.
Two Task Force members strongly feel: Rotated headings are not necessary or desirable in catalogs that offer keyword-in-heading (k-i-h) subject searching. In a k-i-h environment, rotated strings only serve to thwart the collocation offered by the original pre-coordinated order of the subject heading strings in the index, and produce unnecessary duplication of records for the patron. Catalogs should offer rotated headings only to provide expanded phrase-searching access to commonly searched-for terms such as geographical names when they do not offer k-i-h searching.
Some Task Force members agree with the “spirit” of this principle, but think it could be expressed better. Users need a “digestable” amount of information presented to them in a format that is easily read. Sometimes, the use of tables, boxes, and even white space, makes displays more readable, even if using these elements takes up some space on the screen.
Suggestions: display in the document the outline of the screen. Also display more hits, so that the second example can be seen to display more than the first.
Some of the Task Force members would like to see Principles 11 & 12 combined into one principle such as: “Provide Logical, Compressed Displays” then perhaps Principle 15 should be renumbered to follow and the examples could then build on the previous. Other TF members recommend that the two principles remain separate, holding that compression warrants a separate principle because of its importance and because of the complexity of its implementation.

Principle 13 (p.21)

Suggestion: In addition to providing logical sorting by default, users should be able to select, either before or after completing their search, an alternative sorting method, including a straight alphabetical sort.
Perhaps it would be useful to offer the more sophisticated user the choice of advanced sorting such as having headings offered in pure alphabetic order or reverse chronological order. If so, then an example of that should be provided.

Principle 14

Suggestion: have an example here?

Principle 15 (p. 23)

There is a potential contradiction between this principle and Principle 12 on compression. We suggest rewording the name of the principle to: “Do Not Truncate Headings in Uncompressed Displays.” We recommend changing the body of the principle to read, “Never truncate the display of individual, uncompressed headings in a headings display. Always display full headings, including any subdivisions or qualifiers. For guidelines regarding compressed displays, see Principle 12.” One Task Force member feels that: Principle 15 should be renumbered to follow Principle 12 and the examples could then build on the previous ones.

Principle 16 (p.23)

Whether books by or about an author are displayed first could be based on how the user searched. For example, if the user performed an author search, works by the author should be displayed first, but if the user performed a subject search, works about the author should be displayed first. In the case of a keyword search, where it is not clear whether the user is searching for works by or about an author, an intermediate screen should be displayed allowing the user to choose which list they wish to see first when a retrieval set would exceed one screen in display. The display is confusing in Figure 7, Example 1, Alternative in a Web-based catalog. A label reading “Work(s) by:” below the name heading and above the title entries, as in the first part of the example, would help.

Footnote 15 (p. 30)

There’s a typo in the first line: $v ought to be $x. Additionally, the following sentence in footnote 15: “This display probably would not be possible for records using the UNIMARC format” is incorrect and should be deleted.

Principle 18 (p.30)

The title-change example is too easy and makes display requirements look simple to address. A better example might be: Atlantic monthly, and, Atlantic. These went through several title changes back and forth and would make an interesting display. Also, the name of this principle does not correspond to the others in that it is not stated as an imperative. It should be.

Comment on Principles 19-20-21:

One Task Force member suggested that “represented in the catalog” be added to the headings for these three principles. This is really a question of blind references. The titles of these three principles read – “Display the heirarchical relationships between (19) headings and their subject subdivisions; (20) a corp. bod. and its corp. subdivisions; (21) a work and its parts.” You would not want a heirarchical relationship displayed that did not lead to a bibliographic record or records. Then Principle 22 reads “Display the heirarchical relationship beween a classification number and the entire classification.” Although the pattern of these four titles is the same, there is a considerable difference between the 19-20-21 group and Principle 22. 19-20-21 should deal with what the library (or range of libraries) has. 22, on the other hand, concerns a theoretical subject structure, and looks beyond holdings. This difference should be brought out in the titles of these Principles.

Principle 19 (p.31)

The example used seems to us to be too simplistic. An added author or author/title example would enrich the Guidelines. Perhaps a heading which includes both author and title, like “Shakespeare, William, 1564-1616” and its various work headings and subject subdivisions could be used as an additional example.

Principle 22

Does this principle hold true for all classification schemes? For example how do the SuDocs or Swank schemes fall into this principle, examples should be included, not just of LC and Dewey classifications

Principle 24 (p.34)

Add this into Principle 5 since highlighting is a good and very common way to emphasize search terms in retrieval displays.

Principle 25 (p. 34)

The Task Force was not unanimous in accepting ISBD display over labeled displays. Concern was expressed about end-user needs since this principle’s only reference to labelling is oblique at best. One Task Force member comments: While I am very much in favor of effective non-labelled (or minimally labelled) displays such as those on catalog cards and bibliograhic citations generally, I think this principle is not well thought out. The IFLA TF needs to think this one over again. I think they should focus on explicitly stating the value of minimally-labelled displays – why they are useful, etc. – and the value of following some standard for such displays. While the principle notes that ISBD is an international display standard and mentions the relation to cataloging codes such as AACR, it fails to note that AACR is a competing display standard. The complexity of this issue is not well presented here. It must be. Another point about this one: the recommendation embedded in the principle – automatically add ISBD punctuation – is not a simple matter (as the footnote to the principle itself makes clear.) Would that mean double punctution in the record with say a pre-ISBD punctuation and ISBD punctuation? Or do they mean not add but replace? Replacing one display standard’s punctuation with another’s may not be worth anything for the OPAC reader.
One suggested re-wording: Take out the phrase “and labelling” in the first sentence. The mention of labelling with no labels in the adjoining figure 13 is confusing, so drop the phrase “and labelling.” Instead, add a footnote that says something like, “For guidelines regarding labelleddisplays, see General Recommendation 14.d.”
Another suggestion: Whether labelling is addressed specifically here or only in General Recommendations 14 d and e, some good examples of both bad and good labels would make the document better.
In Figure 13, the word “proper” is a trifle pedantic. Perhaps “typical” would be better?
Sustained ISBD order offers predictability and intersystem consistency in displays, while a rotated order reinforces the user’s search by highlighting the term(s) actually searched for by him or her. (To address the latter, what about putting searched-for terms in bold in the display regardless of order, as many databases do?) Also, the name of this principle isn’t stated as an imperative. It should be, to follow suit with the others.

Principle 26 (p. 34-5)

The section on “Principle 26, Make the Default Single Record Display the Full Display” is remarkable because it also includes recommendations for children’s libraries. Children (and/or their parents) deserve the same important record display information as is available in larger and/or research libraries (with the option, as stated, that human editors edit cataloguing records, “rather than relying on the arbitrary dumping of fields by a computer algorithm”). However, this principle is targeted for monographic (i.e. book) formats and needs to also address the problems associated with other formats, especially serials and electronic resources, Additionally, some libraries, particularly those with “location complexities,” might need to use a somewhat abbreviated display as a default in order to make location information more readily visible to the user. The note could say that this approach should only be used as a last resort. Full displays are particularly problematic with long serial records, which is doubly troubling because holdings information is so important to serial users.
Electronic resources need “hotlinks” (via the 856 MARC field) to appear prominently in the bibliographic display as well and not at the end of the record.

Principle 30 (p.35)

This is the only principle that has no commentary and no examples. Some others don’t have examples, but they have a bit of comment. This principle is very important and should have a paragraph that justifies the principle. Examples of bad displays and good displays of diacritics, special characters, non-roman, and bi-directional scripts should all be included.
A favorite bad example of diacritics is the difference between the spanish word anos and the word años. The former is the plural of anus, the latter is the plural of year. So the novel Cien años de soledad displayed as Cien anos de soledad is at best an amusing mistake. Our OPAC displays should not do this.
Embarrassing examples aside, principle 30 is weakened by having no supporting rationale or examples.

Editorial Suggestions for Recommendations

General points
The Task Force feels that the organization of the Recommendations needs improvement. (see A.1. General points: e: Numbering of recommendations). All of the Task Force members felt there was difficulty in the way principles did not directly link to the recommendations (and vice versa). One Task Force member suggested incorporating the recommendations within the appropriate principles, i.e. state the principle, then an example, followed by recommendation(s). If the document could be organized this way, some of the more lengthy examples could perhaps serve both for principle and recommendation.
One Task Force member comments: I think the recommendations must be explicitly tied to the principles they are based upon. The paragraph on organization in the preface to the Guidelines, states that the recommendations are “meant to be a more detailed expansion of the principles into actual practice.”
I also think that one method of organizing the principles and the recommendations should be used and not the two different approaches the Guidelines draft uses. There are good reasons to choose either (and good reasons against each), I just think they should choose one and stick to it. Then index or x-reference the principles and recommendations. I like the suggestion that a table or graphic be used to show other groupings of the principles and recommendations.
Although I like the general to specific approach used in the principles section, I don’t think it works very well--principle 30 seems very general to me; a revised principle 25 on display standards would also be general not specific.
The organizing rule for the recommendations states that it will use the 4 main types of searches used by readers plus a section on general recommendations, but it doesn’t seem to actually do this. The recommendations are organinzed into one general group and 7 specific groups. But I like this second approach better than the first. A section for the most general statements (principles in one part of the Guidelines, recommendations in the other), followed by some number of specific groups. I think this is likely to be more like the 7 the Guidelines actually use than the 4 they said they would use. But it needn’t be exactly those 7 or in the order they are in now.
Comments on individual recommendations

p.36, General Recommendations #1

One member asks: Although the Introduction to the Guidelines (Intro p. 5) lists four main types of search addressed in the document, only the first three are listed in Recommendation 1 (a-c). Shouldn’t #4 “A search for works that take a particular disciplinary approach” also be there? Also, consistent with this Task Force member’s comments on Principle 5 (expanding to allow “other search terms”) could the Committee add wording that would allow other displays of those other search terms? Does the wording as it is now allow for, say, a numerical display of all issues of a series title? Could it accomodate a search for a map of an area by scale, or coordinates? This Task Force member would like the wording to respond to the wide potentialities of online searching. Perhaps add: “The display should be able to accomodate the searching capabilities of the system.”

p.36, General Recommendations #3

One Task Force member comments: Should references be selectable or transparent? In the electronic environment, it might be better if they are semi-transparent: users are automatically taken to the “destination” pointed to by the reference, but are notified during the process that they are being redirected, much like a URL “redirect” page on the Web. If a user is searching for something and the only thing they find is a reference, why wouldn’t they choose to follow it?

p. 36, General Recommendations #8

In order to get the benefits of both a standardized display and user customization, the screen should have a reset-to-default-display button visible at all times. (This feature would work similarly to a reset button on a TV monitor.) This would allow public services personnel or users themselves to easily get back to familiar territory with a display.
One Task Force member comments: I strongly disagree with the statement: “Displays should be the same no matter how the user has accessed the OPAC . . . through every kind of system module, such as cataloguing, circulation, or acquisitions modules”
While I largely agree with this principle for default public displays and for sorting/filing rules, I disagree that technical staff should be forced to have the same default display as end-users. Catalogers and other technical staff work with records in a very different way than end-users do and benefit greatly from displays that are customized to their needs, which primarily involve activities such as editing records (either individually or in groups), comparing records to one another, comparing headings fields in records to headings fields in authority records, etc.
I also feel that end-users should be able to control some aspects of display, and that libraries should be able to set defaults differently for different user groups as they see fit (adults vs. children; English vs. Spanish, etc.) Perhaps a better recommendation would be that libraries should be able to control the level of customization that is available in their catalogs depending on the method of access.

p. 38, General Recommendations #11

The reference “also note the exception to this rule below under Names” is vague. It should refer to a specific recommendation (a single numbered sequence would facilitate this) and/or page number.

p. 38, Figure 15

Why does this display begin with line no. 2. Shouldn’t it be 1?

p. 48, letter g

The second sentence is hard to read with the (nested!) parenthetical expressions. It would probably be easier to understand if the info in parentheses were pulled out into a following sentence.

p. 51, #1

This recommendation and many, many that follow it refer to “the principle of sorting elements.” The reference should specify which principle by number if not also page number, to make it easier for the reader to track it down. The Guidelines should not assume that the reader will read them from beginning to end, but will rather dip in here and there as interest dictates. To accommodate that kind of reading, the Guidelines’ internal references should be consistent and specific.

p. 53, #1 under “Display of the works under . . .”

This recommendation is complex enough that it warrants an example to clarify just what is meant and how the display should work.

p. 54, #6

The reference “should be the same as that described under WORKS” should be made more specific. There are other identical references that follow (e.g. p. 59, #5) that should be changed, too.

p. 55, #1 under “Display of manifestations . . .”

The Task Force recommends that re-sorting and limiting be addressed in principles. Once that is done, this recommendation should refer to the principle on re-sorting.

p. 56, footnote 26

Make the reference to the principle on compression more specific.

p. 57, #3

The Guidelines don’t mention much, if at all, the use of form or genre headings under thesaurus control. It would be useful if they did. In this recommendation and related figure, the domain could be expanded to include examples of form/genre headings that are labelled with the thesaurus from which they come. The use here of LCSH and MeSH is too limiting considering the many thesauri that a catalog could implement.

p. 61, first paragraph

The recommendation mentions browsing the shelflist. We suggest there be a separate recommendation in this section on how browsing class nos. should work, that it be bi-directional, for one thing. The general recommendation on browsing the shelflist should also say (1) that users should be able to display an entire classification system, i.e. letters or numbers with captions, at the highest level of its hierarchy, e.g. for LCC, A General Works, B etc., C etc., (2) that those levels should then be able to be expanded through the hierarchy down to the individual class number, (3) that users should be able to move up and down through the hierarchy, and (4) that users ultimately should be able to browse class numbers and their corresponding bib records bi-directionally. If the Guidelines are recommending hierachical displays, this feature should be possible, thus allowing the user to visualize the classification scheme as a whole, to drop into it in an area of interest, and then to browse individual numbers and the records associated with them.

p. 61, #3 (at the bottom)

This recommendation doesn’t seem to differ much from Principle 22. If it is to be retained, it ought to refer to the figures that we are recommending be added to follow Principle 22.

p. 62, “Display of classification number in a single bib . . .”

The first sentence could use an example to illustrate it. It’s not obvious or intuitive how what is being described would be done.

p. 64, #2 (under “Display of multiple records retrieved . . .”)

This recommendation mentions re-sorting records in a retrieval set. If a principle or recommendation on re-sorting is added, this recommendation should refer to it.

Glossary

As we have stated, it would be helpful if terms that are defined in the glossary (as opposed to defined in the footnotes) could be highlighted in the text of the document. We recommend adding the term “syndetic structure” from pg.15.

Glossary Terms

Heading

At the end of the first sentence, the phrase “or with other headings matched” is rather opaque. The IFLA Task Force should try to restate to put across what they really mean.

Qualifier

In #2, there finally is a general description of limiting/filtering. As we’ve said many times already, this should be put into a principle, to which this definition could then refer.

Summary display of headings matched AND Summary display of bib. records

Wouldn’t a display of headings re-sorted by date still be a summary display? The use of the word “alphabetical” is problematic here. Alphabetization may be rightly recommended in the Guidelines as a default, but it is not a criteria for determining a summary display.

Respectfully submitted: 4/15/99