Final Report of the AISC Indexing Task Force
November 2002

Introduction:
The Indexing Task Force strove to develop indexing that balanced the current practices in the 2 systems and that also included the indexing guidelines of the Z39.50 National Profile, Functional Area A, Level 2. We ran into a few snags while trying to comply with Level 2. Although the indexes appear to be set, the tags to be included in each index are not yet mandated. In addition, the distinction between tags that ought to be included in keyword searches vs. headings/browse searches is not clear. We could not comply completely with guidelines and standards that have not yet been accepted. However, catalogs appear to be judged as "percent compliant" based on a number of search attempts, so our product does not have to be perfect to be considered acceptable.

Indexing Task Force Assumptions:
To the best of our understanding at present, the Giant Union Catalog (GUC) will not be a giant catalog with copied bibliographic records and its own index. Instead, the GUC may be an "index of indexes" or a "pass-through" in which a search request in the GUC is sent to the CC and SUL union catalogs, and results are combined into one retrieval set. However the GUC winds up being structured, the Indexing Task Force felt safe in basing its work on the following assumptions:

  1. The GUC will be a subset of the lower catalog indexing. A tag cannot be indexed in the GUC unless it is indexed at a lower level.
  2. Tags may be indexed in specialty indexes at lower levels, but found only in a general index at the GUC level. For example, the GUC will put all 69x (local subject) tags in the "Subjects,All" and "Subjects,Other" indexes. The GUC cannot attach a specialized label to 690 because local practices vary, and 690 does not indicate the same specificity in each catalog that uses the tag. However, the University of Florida catalog will contain a specialized index of 690:9 and 691:9 for "Subjects, Reversed Georgraphic" because those tags and indicators are reserved for that specialized use.
  3. Although the group agreed in principle that the GUC should be weighted towards precision rather than recall, at the end we included more tags/indexes in the GUC on the assumption that if these indexes are planned to exist in the CC and SUL union catalogs, there will be no substantial increase in costs (indexing time, storage) to include them in the GUC. We also kept in mind that the defined tags/indexes for the GUC need not be visible options on the public pulldown search options.
  4. Since the GUC will be a subset, its indexing will change as changes are made in the CC and SUL union catalogs.
  5. Although local and specialized indexes will be created at lower levels, the general indexes and the tags/subfields chosen must remain consistent across the CC and SUL indexing. For example, the 100k tag/subfield cannot be included in the Title index for the CC but in the Author index for the SUL. Changes to the proposed indexes may (and probably should) occur, but decisions must be acceptable to and implemented by all parties.

We recommend not indexing the 034 tag (Numeric Cartographic Data), the 043 tag (Geographic Area Codes), and the 052 tag (Geographic Classification) until our profession has established new standards for treatment. Codes and numbers of fewer than 3 characters cannot be retrieved efficiently by Keyword Indexes; they would likely need to be retrieved through the creation of look-up tables, which can be designed more effectively under established guidelines.

CCLA and FCLA staff will work cooperatively on elements that support the indexing functionality but whose technical details do not require consensus by the Task Force, such as normalization routines.

As groups define additional indexes for the CC and SUL union catalogs, alterations to the base indexing may be recommended. Adjustments may be made as long as they are communicated and agreed-to by all parties.

Click here for the latest Word document that contains several charts. The Headings/Browse chart and the Keyword chart are arranged in tag order, and detail the proposed index and corresponding subfields for that tag. The current FCLA and CCLA indexing is included to serve as a point of reference for anyone who wishes to compare proposed vs. current. The Index Codes and Names chart lists all the indexes and provides a description of the codes. The tags on these charts are included if any subfield within that tag is included in that index. For example, tag 520 is listed in the URI index, but if you consult the Headings - Tag Order chart, you will see that only $u (the URL) portion (if present) in a 520 tag will be indexed in the URI index.

Page updated: 25 November 2002


 
 
Navigation Bar