HOLLIS Administrative Advisory Committee
Meeting 55
February 2, 1993

1. Report from the HAAC Task Force on a Facility to Enter Through the Catalog in HOLLIS End-User Requests for Retrieval from Storage (FETCHERRS), continued
2. Preliminary Report of the HAAC Task Group on Large Index Searching in HOLLIS (LISH)

Attending:

Bob Buckwalter, Marianne Burke, Priscilla Caplan, Michael Fitzgerald, Dale Flecker, Lynne Schmelz-Keil (chair), Carrie Kent, Mark Van Baalen, Judy Warnement, Robin Wendler (recording), Charles Willard

Absent: John Collins, Rodney Goins, Ted Pappadopoulos

Guests:

Jennifer Hanlin, Charles Husbands, Ellen Isenstein, Curtis Kendrick, Barbara Mitchell, Sue Parker, Naomi Ronen, MacKenzie Smith, Dorothy Solbrig

Future Meetings: In the OIS Conference Room:

                    February 22, 3:00-5:00
          March 17, 10:00-12:00    
          April 13, 3:00-5:00
          May 18, 3:00-5:00   
          June 15, 3:00-5:00

1. Report from the HAAC Task Force on a Facility to Enter Through the Catalog in HOLLIS End-User Requests for Retrieval from Storage (FETCHERRS), continued

Dorothy Solbrig distributed a page of revisions to the report, including 1) change to the circ display, 2) provision to include the delivery date and time, and the phrase "press ENTER to complete your request, and 3) strengthen the request for the "wants" function. Cilla Caplan asked how many different delivery schedules HD supported, and Curtis Kendrick tallied it at six at this point. Cilla noted that this would require at least an extra system table, and would become much more complicated if the system were to allow recall to other than the holding library. There was agreement that any delivery date and time supplied to users must be accurate and reliable. The revisions were accepted in principle.

The group than returned to last meeting's discussion of the report, focussing on Mark Van Baalen's question: how does the user know how to proceed? They considered trying to explain the GET command in the options area, e.g. "Use GET to request starred item", or adding the text "At HD. Type GET *1 to retrieve" to every appropriate item line in a circ display. The group discussed the balance between the need to provide detailed explanations for novice users and the need to facilitate access for experienced users. Barbara Mitchell felt that there are many library activities which are standard, which library users can be expected to know or to learn, and once learned, need not be prompted. Carrie disagreed, stating that not everyone is familiar with many activities which library staff consider central functions. Ellen Isenstein pointed out that SSHUSH intends to review use of the options area soon. Charles Willard wondered if it would be possible to give the user a choice of experience levels for using the system. HAAC recommended that the FETCHERRS committee model the options and survey some HOLLIS users for their reactions, being careful to include users at those libraries which do not currently use HD.

Carrie also expressed interest in changing the circ screen message for a requested item to indicate that it is in transit. Charles Husbands wondered how quickly the change would have to appear; this would be a major change to HOLLIS, because at this time nothing a catalog user does changes any HOLLIS files. Dale concurred, stating that this change carries a significant cost: it means OIS would be unable to close files to updates in order to perform system maintenance. Another consideration is that not all requests are handled through this GET mechanism, and it might not be possible for all request transactions to be reflected. The group identified many anticipated developments which would require changing HOLLIS from a read-only system to one which could be altered by patrons, such as patron-initiated recalls, holds, and renewals, ILL, requests for CRL materials, and rush processing requests. The difficulty will be in identifying the right moment to move toward patron-update in HOLLIS. The question requires thorough investigation.

Many parts of the FETCHERRS report require further functional or technical specifications. OIS will identify those issues and reconvene FETCHERRS as necessary.

HAAC approved the report, and thanked the group for their hard work on this topic.

2. Preliminary Report of the HAAC Task Group on Large Index Searching in HOLLIS (LISH)

Michael Fitzgerald summarized the charge to LISH as to consider the effect of the retrospective conversion project on index search and retrieval in HOLLIS, and to identify ways to ameliorate the negative impact of greatly larger files on intellectual access. The group was charged to recommend solutions which would be economically and technically feasible within the current HOLLIS structure.

The HU database is expected to grow to over eight million records in the course of the next six years. There will obviously be larger files under headings. The largest impact will likely be reaching the maximum number of hits more often. Therefore, the task group saw increasing precision as the most pressing need. They made a variety of complementary recommendations:

  1. Make qualification available in keyword indexes
  2. Add a "system working" message during long searches
  3. Provide a BREAK key to terminate long searches
  4. Consider pre-selected subset files, such as English language only, or the latest ten years only, to reduce both large numbers of hits and to reduce computing resources necessary for many searches
  5. When the maximum hits are reached, provide an automatic screen which encourages the user to fill in qualifying information and reissue the search
  6. Raise the maximum hits in the keyword index from 250 to 1000
  7. Provide ability to reissue a string search, eliminating authority references
  8. Reinstitute the personal name/title index
  9. Expand qualification by LOC code to all string indexes.
  10. Provide ability to convert a no-hit string search to a keyword search
  11. Revise HELP to provide two levels of support, hints and detailed explanations
  12. Review error messages to provide more useful messages when possible
  13. Provide exact-word symbol for string indexes (to disable automatic truncation)

The group also identified issues for further discussion, including the use of sublocation information in searching and the possible change of the default Boolean operator in keyword searches from AND to ADJACENT. Carrie noted that the Boolean operator change could result in a long, painful transition for users, and would require very careful thought.

HAAC members agreed that the ability to qualify keyword searches was a top priority. Dale reminded the group that NOTIS will not release the source code for the NOTISEARCH software we use for keyword indexes. Therefore, it is difficult and in some cases impossible to modify the ways it performs a certain function. The need for qualification may require us to investigate other keyword searching software. In NOTIS, rather than constructing a set of matching titles and then eliminating those which do not meet the language or date requirements, each language code or date is treated as a word, and the system creates the set of all titles in English, for example, and then looks for overlap with the sets created for each other element of the keyword search. This degrades response time. There are a few new keyword search engines on the market, but OIS has not investigated them. Lynne Schmelz-Keil suggested that OIS pursue this. Cilla Caplan felt that it might be possible to provide some of the desired functionality with the NOTISEARCH engine, and that this would not preclude us from investigating other available software.

Cilla also noted that we should test and monitor the impact of increasing the maximum hits limit.

Marion Schoon reported that users may be put off by the use of the term keyword rather than just "word," which is commonly used in other systems. She also mentioned that unfortunately HOLLIS treats the word "keyword" followed by a search term as an error.

Dale mentioned the option of loading a copy of HU on another platform, where sophisticated, resource-intensive searching would not impact response time in the HOLLIS system. One disadvantage of a copy would be that system integration would be lost; the copy could not be maintained in synch, and updates would be reflected periodically.

HAAC will continue discussion of the LISH report at its February 22 meeting.

3. Minutes of previous meeting

The minutes of the December 21th meeting were approved with minor changes.


Top

Return to the HAAC Home Page

rev. 8/8/96