This discussion applies to situations where institutional policy permits both the single-record approach and the separate-record approach to cataloging remote access computer file serials (as described in CCM module 31).
Publications which exist only on the Web range from counterculture zines to scholarly publications (e.g., Salon [http://www.eclectica.org/v2n5/salon.html] and Slate [http://www.slate.com/]). These publications originated online and obviously must be described in records for the electronic version, since that is the only version extant. Whether popular culture zines or newsletters are cataloged depends on the selection policies of the library.
Note: The IFLA functional requirements describe four layers of hierarchy: work, expression, manifestation, and item. While "manifestion" refers to PHYSICAL differences, "expression" refers to differences of CONTENT. "Equivalent manifestation" is used in this section as short-hand for "physical manifestion with equivalent content." The discussion below seeks to clarify what constitutes "equivalent content" in general. Sections B1-B3 discuss some specific situations.
Many publishers currently publish electronic versions that have identical article content as their print counterparts. The majority of these journals are issued by the original publisher, often in publisher-based aggregations (see below) or in preservation-type projects (e.g., Project Muse, JSTOR). They are the best candidates for single-record treatment.
When presented as scanned images or PDF files, the cataloger can usually assume the content has not changed. This judgment should be made on the appearance of the journal articles themselves; do not consider hypertext links, home pages, splash screens, advertisements, (or their absence), licensing information, and other perfunctory pages as additional content sufficient to merit separate records (Related site, Category D below).
As previously stated, omission (or inclusion) of commercial advertising, incidental illustrations, reader comments, corrections, time-dependent announcements, and other such ephemeral data is not sufficient cause to consider the online version to warrant a separate record. In contrast, however, evidence of the omission of articles would make the electronic version an unsatisfactory substitute as a source for an article cited in an index or footnote. Therefore, such content differences require that the online title be considered a related electronic serial ( Category C below).
Here are some rules of thumb when comparing the content of two versions:
(Comment)
Aggregator databases are large databases that aggregate, or bring together, the full text of journal articles. In most cases, the journals have print equivalents. The aggregators fall into two broad categories: those that aggregate publications of a single or small group of publishers and those that are subject-based and include articles from a vast array of publishers.
Most publisher-based aggregators, such as Academic Press, American Mathematical Society, Springer-Verlag, Wiley, retain the concept of "issues" so the user can locate a particular volume and date. Illustrations, tables and photographs are usually included. The publisher may offer a search engine to locate articles by author, title, or subject. We believe this type of publication is well-suited to the single record approach.
Subject-based aggregators are usually presented as a searchable database, such as the ISI fulltext databases (e.g., ABI/Inform) or Lexis Nexis. Usually these databases contain fulltext or selected articles, with illustrations, brief articles, syndicated and wire-service articles, and other features omitted. The concept of "issues" no longer exists. Since there is no chief source for the component journals, a bibliographic description cannot be prepared according to the traditional library model. The single-record approach can be used for this type of aggregator service, but a user notification should be included saying that only selected article text is available online. At this time, the PCC Task Group on Journals in Aggregator Databases has focused on developing a schema for describing electronic journals embedded in subject-based aggregators and a demonstration project distributing vendor-supplied records for them. Naturally these titles will be cataloged separately from the print but data elements may be included (e.g., OCLC numbers of the print) so that individual libraries may merge them with their print records should they choose to do so.
Publisher-based aggregators could also be cataloged in this way, although it seems unlikely that the publishers would be able/willing to supply AACR2r MARC records for their titles
Some publications originate in one format but have since discontinued that format in favor of another. The most common example seen of this is a print publication discontinued in favor of an online version, but other possibilities exist. The later version may have some overlapping coverage with the original version (ALCTS newsletter/ALCTS newsletter online) or may have no overlapping coverage (CONSER /CONSERline). The second case clearly is a change of format and separate records are recommended (especially in light of the fact that the numbering will often not follow the pattern of the original version). The former is more problematic as the overlapping issues could be considered an "equivalent manifestation (see below) while the current issues could be considered a change of format. (Comment)
For example, some issues of the print version of LITA newsletter also appear online. However, current issues of LITA newsletter are ONLY available online.
In this case, the Committee recommmends using separate records for the print and online versions, but using single-record conventions to describe the existence of overlapping issues (see OCLC #5757570, OCLC #31406418).(Comment)
Separate records should be used to describe publications that have substantially different content, such as Library journal and LJ digital or the New York times and the New York times on the web. Since these versions are so different it is not appropriate to make a 776 "other physical form available" note. Instead, use a 775 "Other Edition Entry" or a 787 "Non-specific Relationship Entry" and an 856 42 related publication link. (Comment)
"Related Sites" includes two distinct categories of online resources: sites which provide partial information related to the e-journal; and other related resources.
Some related sites are appropriately cited by the provision of an 856 41 $3 in the print records (CCM 31.16.2), but generally do not merit their own records. These include sites containing indexes, tables of contents, abstracts, selected articles, supplementary information, or updating information. Most such sites do not need to be cited if the journal itself is available online, although supplementary or updating information available at a different site from the journal could be useful. (Comment)
An interesting cluster of resources to consider with respect to Print serials/Related electronic serials/Related sites is the title Organometallics (print: OCLC #7411854), Organometallics (Online) (OCLC #39280641--http://pubs.acs.org/journals/orgnd7/index.html), and Electronic Supporting Information (http://pubs.acs.org/subscribe/journals/orgnd7/supmat/index.html)
Resources in this category are cited using an 856 42 field in the print record. One example of a related resource would be a web site for an organization included in the record for the group's newsletter.
Many publications cataloged as serials in paper format are now on the Web in database format. These range from publications that were issued as frequently revised editions, e.g., the NASIG membership directory, to ongoing reference publications, e.g., Handbook of Latin American studies. According to current USMARC policies for content designation, records describing such resources are designated "monographs" (Bib lvl m). This can be confusing to users who are accustomed to looking for them in serial union lists and catalogs. Treating them as monographs also takes them out of the CONSER domain. One can argue that, since they are now monographs, they must be cataloged separately from the paper version. Or one can argue that, in the interests of keeping them in the serials arena, they should be cited on the single record (either as an equivalent version or as a related web site, depending on the content). Interestingly, ISSN and NLM have recently experimented with treating such databa ses as serials and ISBD(S) seems to favor such an approach.
GPO catalogs all versions of a publication using the single record approach whenever more than one version exists. Current cataloging now includes the following note: Internet version also available (content may vary from the physical version and/or the content of the electronic version may have been changed by the publishing agency). GPO also does excellent maintenance, keeping PURLs up to date, noting sites no longer available, etc. Rather than create separate records, it is strongly recommended to accept GPO's cataloging. There are, of course, many government serials that exist only in electronic format, for which GPO does not use a single-record technique.
Sites which generally do not warrant cataloging include sites with subscription information only, sites for the associated body only, and sites with advertising only.
A single record is recommended in any of the following situations
A separate record is recommended in any of the following situations
Single record: When a patron is looking for a specific issue of a serial, a single record with the holdings for each version lined up together provides welcome one-stop shopping. The risk of a user overlooking the correct record are much less when there is only one index entry per title.
Separate records: When the patron wants to limit the search to electronic versions, a separate entry for the electronic version may be easy to spot in the index due to display of the uniform title (e.g., LITA newsletter (Online)) or the GMD (e.g., LJ digital [computer file]). But if they don't find the issue they need in electronic format, will they remember to return to the index and look for other formats? Also, occasionally the two versions are widely separately in the index, particularly if there are many monographs with the same title filed between them, e.g., Blood [a serial], Blood [twelve monographs by that name], Blood (Online)
Single record:
When no copy is available, describing electronic versions from scratch is time-consuming in contrast to using existing print records and adding 3 or 4 standard fields. Often the addition of the extra fields is identical for each title within a publisher's service and can be done by nonprofessional staff or with constant data records (or their equivalent).
A library that includes federal government documents in its catalog will want to incorporate GPO single records rather than creating original separate records and keeping up their maintenance.
If a library is responsible for the digitization, it is in a good position to judge whether the content is equivalent. If no additional features are added, generally a single record will be much less expensive. On the other hand, some may feel that the maximum amount of detail needs to be included on the record.
Separate records:
Libraries that routinely recatalog the print record (including consulting the printed piece) when making a single record will not realize any savings. Likewise, cloning the single record as a basis for the separate record may be very time-consuming and involve lots of decision-making, such as upgrading to AACR2r, or determining which notes, added entries, and linking fields apply to the electronic version if it covers only a portion of the time span of the paper. The opposite approach is exemplified by the California Digital Library consortium which uses the print record regardless of whether any UC library owns it.
If separate records are available for copy cataloging or for sale, it will be more economical to simply load them into one's catalog than to search for print records and add version information. This is the hope of the Aggregator Task Force. (Comment)
Single record: When using a single record, the library must determine whether to add its holdings symbol to OCLC for the print when they only have access to the electronic version. This would waste time and effort on unfulfilled loan requests.
Separate records: Most license agreements forbid lending of electronic titles. The few who do allow it often have extremely cumbersome paperwork obligations. Many universities have a blanket policy against lending electronic titles. Is it appropriate to add the library's holdings byte to a record for a serial you cannot loan?
Single record: If records are received from an outside agency, the single record technique requires manual intervention to identify and replace/update the proper record. If the agency furnishes single records, like GPO, it is surely worth dedupping to avoid creating original records.
Separate records: If cataloging copy received is for separate records, most libraries should be able to add them to their catalogs easily.
Single record: When a library receives both the paper and electronic versions, they are often alerted to changes in one version by changes in the other. If both undergo the same change, the cataloger only has to adjust one record. If one can obtain revised records, such as those from GPO, maintenance is expedited. Other changes depend on link checking, consulting "what's new" pages of electronic publishers, and checking for retrospective coverage. Deletions are time-consuming to process when each field pertaining to the electronic version must be stripped manually.
Separate records: If revised separate records can be obtained through vendors or fuller-record-notifications, it is a strong reason to utilize separate records. Otherwise one depends on link checking, consulting "what's new" pages and checking online. If the library receives another version, often the cataloger must perform redundant maintenance and there's a risk the records will get out of synch. Deletions of single records are easy and quick.
Single record: The already-lengthy records grow by 3-5 more fields when electronic data is added. Already some serials records are too long, especially those with multiple 510 fields.
Separate records: Electronic records with separate fields for type of file, system requirements, mode of access, and access limitations may not be appreciably shorter than the original print record, especially if call numbers, subject headings, 510 notes, bibliographic history and added entries are transferred from the print record. If a library collects a significant number of electronic journals without discarding the print, the catalog database may grow exponentially, perhaps incurring expenses to increase its capacity.
Single record: If two versions were represented on a single record and the print ceases, should the bibliographic record be left open? If the record is closed, will the user understand that the electronic version is still available?
Separate records: If a separate record overlaps with the print record, should the user be directed to a successive record for the very same resource described in the single record, with the very same 856?
Many of the comments received relate to specific integrated library systems. Hypertext links, holdings displays, statistical programs, and loader programs vary widely and strongly influence perceptions of what is confusing or efficient. Likewise the talent, experience, and level of staffing make some institutions fearful of exercising independent catalogers' judgement, while others expect to do that with any serials cataloging.
Text written by Crystal Graham; revised by CONSER Working Group members Steve Shadle, Cecilia Sercan, Valerie Bross.
990414 CONSERWG; last updated 990601