Bill Anderson, Matthew Beacom, Liz Brown, Dale Chatwin, Anne Daniels, Crystal Graham, Beth Guay, Les Hawkins, Jim Holmes, Wayne Jones, Hien Nguyen, Steve Shadle, Pamela Simpson, Allison Sleeman
John Attig Winnie Chan, Laurel Jizba, Patricia Luthin, Dave Reser, Sally Tseng,
AACR Group 2B was charged to draft rules to be added to Chapter 9 in order to accommodate cataloging of electronic serials. Originally this included recommending rules governing the basis of the description for direct and remote access electronic serials, "considering the relative inaccessibility and/or volatility of descriptive data" for these serials. The original charge also included recommending rules to accommodate the description of non-sequentially issued electronic publications and electronic publications without a chronological or numeric designation that are clearly intended to continue indefinitely.
After some discussion throughout early 1998 and summer 1998 ALA, we decided to focus on remote access resources rather than trying to handle treatment for both direct and remote access as suggested in the original charge. We amended our charge to include:
- Recommending appropriate cataloging treatment for online resources, without being constrained to drafting specific rules for specific chapters.
- Conducting a study of online resources and making recommendations based on data gathered. By looking at the nature of the resources and collecting some demographic information, it was hoped we could better make observations and recommendations about title change situations and selecting appropriate source of title.
Additionally we were also asked to consider the incorporation into AACR some practices that have been documented in the CONSER Cataloging Manual Modules 31.10 and 31.4.3 respectively:
- Use of the 362 1 "Coverage as of" note
- Use of the "viewed on" date in source of title notes (and coverage as of note)
Random samples of 100 titles each were selected from the CONSER database and the most recent ARL directory. We assumed the CONSER sample would represent items most likely to meet the current AACR2 definition of serial and that the ARL sample would contain at least some resources that may not so strictly meet that definition. We hoped this mix would allow us to compare resources that don't fit the current serial definition (databases, web sites, etc.) with other electronic serials.
A data collection form (http://lcweb.loc.gov/acq/conser/group2b/surv-form.txt) was developed and volunteer data collectors from the group were assigned titles to assess. We were able to attempt a connection with about 82% of the sample resources. A summary of selected study results is available from: http://lcweb.loc.gov/acq/conser/group2b/contents.htm
The total number of confirmed title changes in the combined CONSER and ARL sample was about 12% (combined number from tables 3 and 4 of the summary document). We assume this percentage will be higher in the future as data collectors identified a number of online versions of print titles in which the publisher had not yet digitized issues far enough back to encounter a known print title change.
In terms of how title changes are represented online, there were no real surprises. We found that publishers sometimes maintained the identity of earlier titles along side new titles at a single site and some publishers of serial and non serial resources completely reformatted earlier titles leaving little trace or explanation of the earlier title. These two approaches were used about an equal number of times among the total confirmed number of title changes (including items that meet the current definition of serial and those that don't). There were a few cases among this group where it was unclear what was happening at the site. Among the resources that could be said to clearly fit the current AACR 2 definition of a serial (table 3 of the summary document), we found that the first approach was used a little over half the time (7/13), the second about a quarter of the time (3/13).
In general the group agreed with provisions outlined in AACR Revision Group 1 draft report to use successive entry cataloging for "non-integrating" electronic serials and use latest entry cataloging for obvious integrating title changes.
In discussing some of the pros and cons of this preference, group members mentioned several potential difficulties. For integrating entities, it may be difficult to determine exactly when and whether a title change has occurred. For several items in the survey that completely reformatted into a different sort of entity and title, it would have been difficult to know that at one time the there was a different resource, except for the fact that there was a CONSER record or ARL directory listing for the earlier title, describing a different sort of resource. It might not always be clear at the outset that an item is or will be an integrating entity. Will the distinctions between integrating and successive serial publications be clear enough to choose one or the other approach?
Another point made was that for print serials whose online version is an integrating entity, e.g. clearly an article database, there conceivably could be two different cataloging approaches used for the print and online versions. Situations where there are instances of two or more records for a succession of print title changes and a single record for online title changes will need to be addressed by the rules, particularly AACR2 12.7B7. The Group also mentioned potential problems caused by an overly large single record for the online version in these situations.
There also was some concern in general with the fact that two different cataloging conventions would be applied to electronic resources. Some group members favored using latest entry for all serial and integrating networked resources.
A final point regarded cases where a publisher maintains access to some issues which contain the early title and other issues the later title at one Web site. This may mean that the home page presents the title of the current serial. Using successive entry for this situation would mean two records are created which each point to a single Web site showing the current issue. The earlier title would be buried at a lower level and not introduced by a home page showing its title..
Sources of title
The data collection form asked for an opinion from the surveyor about the most appropriate source of title and reasons why a particular source was selected. Though it would not be a good idea to make a recommendation based on the most "popular" source, some useful information was gained from this. The preferred title sources fall into several basic categories, the home or initial page of the resource, an issue or earliest issue, and other available sources which don't include the home page or issue. This information is given fuller explanation in the study summary document, but the reasons a particular source was preferred over another overlap among these categories. The most prevalent reasons for selecting a source regardless of whether it was a home page or an issue include:
- formality of the title presentation (the intention of the publisher to present the title formally) on a given source over others
- the fact that there was no other appropriate source or the title of the item did not appear on another source
In some cases the surveyor indicated that title presentation was formal, clear, and or consistent on all possible sources, so there was really no preference. In other cases, selecting a particular source made it possible to select a fuller title, a more consistent title, keep the title of the online version in sync with the print version or avoid what might have been an unintentional change of title on the publisher's part.
There were certainly a variety of possible title sources available for many of the items viewed. About 90% of the items had a introductory or initial page that showed a title of some sort. About 40% of these pages were referred to by the publisher with a name or stated function like "home page", "welcome page" etc.
The data collection form asked the surveyor to record information found in the html document source title on the initial page. Of the 122 responses to this question, 26% matched the title listed on the initial or home page. About 10% contained the name or variations of the name for the initial page: main page, home page, table of contents etc. The remaining responses contained a wide range of title variants, title with subtitle, corporate names etc.
A number of the resources contained pages which acted as a table of contents for several issues or volumes (76%), and as a separate question, pages that acted as tables of contents for individual issues (56%). In both instances, the title of the surveyed item was carried on these pages in about 90% of the cases. It should be noted, that the responses for both these questions sometimes referred to a single page, sometimes not. It was also not always clear exactly what page was being referred to, whether home page, issue or a distinct separate contents page. The responses to these two questions may be more useful in pointing out that for the resources surveyed, there were a significant number organizing the content using the metaphor of issues and volumes, and associating a title of the larger entity with presentation of designated issue contents.
In about 44% of the items the articles presented the cited title, a variation of the cited title or an abbreviation of the title.
There was some disagreement among group members regarding the appropriate source of title for electronic resources. We generally agreed it was important to select and cite the source of title relying on a formal presentation internal to the publication. In many cases therefore, the preferred source would be an introductory or initial page (home page, welcome page, etc.) for the serial rather than any particular issue. In other cases, where an appropriate introductory page does not exist for the serial being cataloged, an issue or some other internal page would be cited, depending on the circumstance. If a prescribed order of sources were devised, clearly the cataloger would need to be free to cite other sources in certain circumstances.
The group was somewhat divided as to whether a prescribed order of preferred sources should be devised for World Wide Web based serials. It was suggested that the AACR2 ordered specifications of sources (title screen, other formally presented internal evidence, external sources- in chapter 9; and t.p., and title page substitutes in chapter 12) offer the cataloging community a common understanding of what the title is or was considered to be, in a title change situation. Stating a preference for introductory pages to some degree offers a firmer common term for catalogers than the words "title screen."
Other group members were concerned that moving too far away from using an issue or a source close to the issue level (e.g. issue contents page or article/issue) as a source of title separates a primary means of identification from the journal content. It was pointed out that nearly half of the items surveyed contained articles presenting the title of the resource and that article delivery mechanisms (database search facilities) are becoming an important and common means of accessing journal content.
Additional comments pointed out that the services that provide access to many journal titles (e.g. Project Muse and the services of several commercial scholarly publishers) use a common structure, offering a page that is intended to be a primary identifying and title presentation point for a particular journal, and thus a preferred title source.
In general HTML document source title was a much less desirable source of title, because of inconsistency in information presented there.
Use of the "Coverage as of" note and "viewed on" date in source of title notes
It was generally agreed that the "Coverage as of" note and "viewed on" date information in source of title note are critical to the description of electronic serials and serial-like entities and should be incorporated into AACR.
There was one suggestion that repetition of viewed on date be dealt with somehow, that is if a record has both a coverage as of note and a title source note citing a viewed on date, is it necessary to put the date in both?
In answer to this, it was pointed out that the repetition would only occur at the time of cataloging. As the record changes it is likely the two dates will digress as coverage expands backward (as pointed out in the CONSER Cataloging Manual 31.10 (fall 1997)).