HOLLIS Administrative Advisory Committee
Meeting 73
February 15, 1995

2. Report of the Working Group on Cataloging Network-Based Resources in HOLLIS


Bob Buckwalter, Marianne Burke , Michael Fitzgerald, Carrie Kent, Kathy Klemperer, Ted Pappadopoulos, Jon Rothman, Lynne Schmelz (chair), Hinda Sklar, Dorothy Solbrig, Mark Van Baalen, Judy Warnement, Robin Wendler (recording)

Absent: Rod Goins, Eva Moseley

Guests: Dale Flecker, Ruth Haas, Joachim Martin, Janet Rutan

Future Meetings:

     March 22, 10:00-12:00         OIS Conference Room
     April 19,  10:00-12:00        OIS Conference Room
     May 17,  10:00-12:00          OIS Conference Room
     June 21, 10:00-12:00          OIS Conference Room

1. HOLLIS II Update

Dale Flecker described developments in the library system vendor market in recent months. Many more vendors have expressed interest in the segment of the market represented by large research libraries than there were even six months ago, and this is an encouraging sign for us. However, many of these vendors have systems which are still in early stages of development, and their development schedules will have an impact on the timing of our system evaluation process. If we specify our needs too far in advance of the market, we risk having to redo some of that work as time elapses and the world changes too much before we can act on our specifications. An informal group of large research libraries continues to meet to try to define common needs for a next-generation system and to communicate those needs to the vendor community. This group will release a white paper soon. The Automation Planning Committee (APC) will be distributing a paper in the coming month about the process to be used in selecting HOLLIS II.

2. Report of the Working Group on Cataloging Network-Based Resources in HOLLIS

Robin Wendler presented the report of the Working Group on Cataloging Network-Based Resources in HOLLIS. The group concluded that libraries do have the need to be able to catalog some network-based resources in order to make network resources relevant to their collections more readily identifiable and to collocate network resources with traditional resources. The report recommends that technical, policy and procedural changes be implemented to support cataloging of network-based resources in HOLLIS.

While HOLLIS itself is unlikely ever to support so-called "hot- links" from a bibliographic record to a resource, the group recommended that the USMARC 856 field (Electronic Location and Access) be implemented in the HOLLIS holdings format, and that URLs be input whenever possible. Using URLs in the 856 will allow links to be made from these records in other environments, such as HOLLIS II or an alternate intermediate server.

However, the group recognized the significant problems posed by trying to bring largely remote resources under local bibliographic control. Network resources change frequently and substantially. They change in content and coverage; their locations change. In order for cataloging to be useful, it must be accurate. Therefore, the group recommended that libraries which catalog a resource be responsible for actively maintaining the accuracy of the information, and to assist in this, recommended that a report of all such records be generated quarterly and distributed to the responsible cataloging units which will verify the location information. The group also recommended that OIS develop the capability to verify location information automatically to the extent possible.

The group recommended that Internet resources be cataloged under a new general LOC code ("net", which would display in the public catalog as "Networked Resource") in order to avoid implying that a user must go to a specific library to access the item. The cataloging library would be required to add a subfield k to the "net" LOC so that the maintenence reports can be routed to the correct library.

In keeping with other Harvard policies, the group recommended that network-based resources be cataloged on the same record with print originals where applicable and appropriate. It also asked that the Standing Subcommittee on Bibliographic Standards and Policy (BSP) formulate cataloging guidelines for network-based resources and corrdinate an informational meeting for staff on the topic.

HAAC members considered alternate words and phrases to replace "Networked Resource", but could not identify a term they preferred. Kathy strongly felt that paid resources should always be cataloged. While the group did recommend this in the guidelines, the structure of Harvard and the many resources supported by a contributed central pool of funds rather than by a specific library makes it difficult to enforce. Other HAAC members pointed out that in some cases the use of paid resources is restricted to a local population, and that in these cases the resource might be better handled in a different way.

HAAC unanimously approved the report, and charged BSP with formulating the cataloging guidelines, passing them by OIS for technical review, and scheduling a workshop to inform the community about cataloging network-based resources.


OIS has conducted a series of keyword stress tests over the years, most recently in January, to determine whether set size or keyword qualification had an acceptable impact. The most recent tests indicated that there was significant benefit to be gained with a minimal impact on response time by raising the maximum set size. Lack of knowledge about user behavior inhibits firm estimation of impact. We know that users look at up to 50 records on average, but while we know that raising the retrieval limit won't help there, it may help those users who want to look at the earliest 50 records, not the most recent. OIS recommended changing the maximum retrieval size in keyword from 250 to 1000 to gather real impact data. HAAC unanimously agreed, and requested that OIS do this quickly, notify reference staff that it will be taking place and is a test, and report back to HAAC at the next meeting.

The recent stress test indicated that qualification still has significant impact. HAAC will continue discussion at the next meeting.

4) Distributed Union Catalog

Recent information from RLG has indicated that they will be unable to load the entire Harvard union catalog in time for RLIN to be used as a mid-term backup for HOLLIS. Kathy sought guidance about whether OIS should produce a new DUC fiche supplement, or whether the libraries could wait for an copy of the union catalog to be loaded on an alternate local server. The Search Engine Evaluation Committee (SEEC) has been evaluating search engines for an alternate local server for some months and hopes to select an engine this summer and have the server up and running by the end of the calendar year. HAAC agreed to hold off on producing fiche for now, and to revisit the issue in the fall when there is firmer information about the timing of the new server.

HAAC Home Page

rev. 4/30/96