MIT Libraries currently create several different tools for resource discovery:
The Libraries continue to discuss and experiment with other possible methods for resource discovery, such as:
There are multiple inter-relationships among these existing and potential tools. Barton and Vera are currently our two "cornerstone" tools, with complementary strengths that justify the expense of creating and maintaining them as parallel systems at this point in time. Barton is the most effective tool for providing the multiple access points and full description of MARC records, which have become increasingly available for digital resources over the last few years. Vera offers other advantages, such as the quick access of a listing tool, broad subject sorting, and resource management functionality not currently provided by standard library management systems.
The Libraries plan to implement an enhanced navigation program, based on Ex Libris Metalib and SFX software, during this fiscal year. The implications for Barton and Vera are not known at this point, but certainly our environment will look different a year from now and continual review of current practices will be warranted.
Digital products are significant additions to library collections. They provide substantial content and impressive functionality. However, they differ from print resources in ways that impact our cataloging decisions: 1. Since they are usually leased and not owned, they are not necessarily permanent additions to our collections. 2. The content of digital products changes frequently, including additions to content, deletions of content, changes in database structure, changes in urls, etc. 3. Digital resources are immediately available via links from the catalog record; print resources cannot be directly linked to access tools and are sometimes truly remote.
Factors influencing whether a digital product will be included in Barton, and whether it will be represented at the product level, the individual title level, or both:
Cataloging managers are charged with considering these factors as well as possible strategies (see below), consulting with Collections Managers and other public services staff, and making judgments related to the best use of staff resources for the purpose of providing bibliographic access to all of the Libraries' collections.
Keeping or adding URL's on catalog records when e-versions of owned print materials are available is a special situation where practice is driven by a desire to add as much access as possible within the constraints of staffing. For serials, the content behind such URL's (on the copy or on the piece) is checked. If it is full-text content, the URL is added or kept, and holdings information is added to indicate availability. For monographs, the content is not checked; if it is evident from the URL's structure that it is not likely to lead to full text content, the URL is deleted. When URL's are kept, no holdings information is added to the records (currently under review).
Whether to attach links to electronic resources to the records for their print equivalents is a problematic issue. Tensions exist between the labor costs of cataloging and the usability of individual records and of the catalog as a whole.
Piggybacking is less labor intensive up front (that is, in creating records). With exceptions noted below, we are piggybacking both serials and monographs when we have records for print equivalents. It is unlikely we could ever catalog all e-resources with given staff levels if we didn't "piggyback". In addition, public service staff have preferred this approach because the browse indexes are much more comprehensible without multiple records.
Separate records are considerably more labor intensive at the point of record creation, except that they sometimes accommodate the opportunity to load vendor supplied records. With appropriate coding, automatic deletion of records is possible in the event that a product is no longer licensed. In addition, this methodology presents fewer challenges for migrating records to new library systems. While the browse indexes are confusing, as noted above, in other ways usability is better: individual records fully describe the respective formats, and e-resources are always visible from the browse screen.
Two exceptions to piggybacking records when a print equivalent is held have been made to date: 1) Books 24 x 7. We loaded vendor records for monographs and a few serials. In those cases where we hold a print title, the vendor record is a separate record from the record for the print version. 2) ACM. After an initial period of piggybacking, we undertook an experimental use of CORC to create brief records regardless of whether we held the print version or not. The effect of mixing piggybacked records and separate records on the usability of the catalog will need to be monitored over time.
MIT Libraries are currently aggressively pursuing pragmatic strategies for cataloging e-resources, on a product-by-product basis. This is an experimental phase and it will necessitate assessment of the effectiveness of various strategies for access to individual products and the impact on of those strategies on the catalog as a whole. During this period, we will utilize various techniques that fulfill the functions of resource discovery and link to item. Records may be briefer than those required when a record functions as a surrogate (as for print resources). In addition, we will utilize various techniques for managing and deleting records that reflect the volatile nature of the digital products.
Sample strategies for adding records:
Sample strategies for facilitating record maintenance:
Outstanding issues:
See also polices and procedures at http://libstaff.mit.edu/colserv/cat/erescat/erescat.htm
Back to the main Collection Services page.
Microsoft Word version of this document.
Last updated on 2001-07-27 by Gordon Thomas.