01. Introduction
General starting points for the indexes in a CBS MARC 21 database
Multiple libraries
The indexes are structured in such a way that they can be used by more than one library in principle within a CBS database, i.e. by more than one ILN (Internal Library Number). But the structure of the indexes is useful also in the case that there is only one ILN / library. But when in the course of time more ILNs / libraries are going to use a CBS, the indexes are prepared for that.
In this context, we don't speak about ORSs, for which the indexes can be structured in a different way in certain respects.
Bibliographical records
For searching by "content" search keys (so not by searching by PPN a.o.) and for the search results, next starting points are applicable:
- All records on the main level are found, independent from the owner of the record, so independent from the fact whether a library has taken a record in use by adding local and/or copy data.
- Local data are found only by the library which has entered these local data, i.e. the concerning library / ILN.
- Copy data are found only by the library which has entered these copy data, i.e. the concerning library / ILN.
Technical limitations and possibilities
By limitations in the index software, it is not possible to make indexes containing fields on main, local and copy level on the conditions that all records (i.e. of all libraries) on the main level are found and also only the local and copy data of the library / ILN that does the search.
Now (May 2021), the following is possible:
- Separate indexes for the main level in such a way that these records can be found by all libraries / ILNs.
- Separate indexes for the local and/or copy data which can be found only by the library that entered the data.
Authority records - generally used thesauri
Only "general" thesauri are used within the generic CBS structure. I.e. thesauri that can be used by all libraries / ILNs within a CBS database and that are accessible by all libraries / ILNs. The authority records in these general thesauri can be linked by means of a PPN to bibliographic fields on the main level, but too and at the same time to fields on local and/or copy level.
For searching by "content" search keys (so not by searching by PPN a.o.) and for the search results, next starting points are applicable:
- All authority records that meet the search criteria, are found.
- Plus the related bibliographical records which contain an authority PPN link on the main level.
- Plus the related bibliographical records which contain this authority PPN link on the local and/or copy level of the searching library / ILN.
- So minus the related bibliographical records which contain this authority PPN link on the local and/or copy level of other than the searching library / ILN.
In other words, not all links are always visible for the searching library / ILN. This means that the command rel tt (i.e. rel(ated) title records) finds only the records that meet the abovementioned criteria.
Authority records - private thesauri
Contrary to general thesauri, private thesauri are not part of the general CBS MARC 21 indexes.
But when these are implemented, the PPN links from the bibliographic records to the private authority records are allowed only to the local and/or copy fields. PPN links on the main level of a bibliographic record to a private thesaurus are not allowed.
Basic assumptions of indexes
- For the CBS MARC 21 bibliographic and authority records, the indexes are based on both the WorldCat as well as the Z39.50 indexes.
- If the set of fields/subfields used for a WorldCat index and those used for the Z39.50 variant differ, the union of both field/subfield sets is used for the CBS MARC 21 index.
- The search terms (mnemonics like ti, au, etc.) and the search attributes (ikt numbers like 1000, 4, etc.) for the CBS MARC 21 indexes are mostly identical to the ones used for WorldCat and Z39.50.
- If possible, the combination of search terms (mnemonics) and the search attributes (ikt numbers) for the CBS MARC 21 indexes both for bibliographic and authority records are in principle identical for the same kinds of fields and subfields.
- An index may describe more than one level of fields (main, local, copy). E.g. a WorldCat or Z39.50 index contains main and local level fields. This index is part of CBS MARC 21 indexes. But additionally, we have defined separate indexes for local and/or copy level, including new mnemonics and search attributes.
- In principle, we support word indexes only, but exceptions do exist.
- Phrase indexes are supported, but only when they are necessary, for instance in case of multi-word units like personal names, URLs, classification numbers.
- Whole phrase indexes are not supported.
- Special indexes, like those for SCIPIO, are not implemented.
- If needed, special indexes for CBS may be added, but they may never interfere with or replace the official indexes, including mnemonics and search attributes.
- If needed, special indexes for customers may be added, but they may never interfere with or replace the official indexes, including mnemonics and search attributes.
- For PPN links between records, subfield $Z is used. The corresponding PICA+ subfield is $9. When these subfields are mentioned in the specifications of several indexes, it means that a PPN link may be present, e.g. between a bibliographic field and an authority record. PPNs can be searched by using the mnemonic ppn:.
Registrazioni bibliografiche
For bibliographic records the following documents have been consulted:
Registri dell'autorità
For authority records the following documents have been consulted:
Search keys and the punctuation used
In case of word indexes, identical search keys can be used for searching WorldCat as well as a CBS MARC 21 database. In this case a colon will follow the mnemonic, like ‘pn:’ for ‘personal name’.
For the phrase version of word indexes, we will use the identical mnemonic, and complete it by using the character ‘p’ for ‘phrase’ instead of the colon.[1]
Indexes for non-MARC 21 fields, the so-called CBS fields
Additional word and phrase indexes for the CBS fields are also part of the CBS MARC 21 configuration.
Search keys
- The search keys for fields based on MARC 21 consist of 2, sometimes 3, 4, 5 or 6 characters.
- The search keys of the CBS fields consist of at least 3 characters.
- Search keys (mnemonics) not mentioned on the OCLC Index Documentation webpages, consist of 4, 5 or 6 characters. This applies to all levels: main, local and copy level fields.
- The search keys for local indexes start with the search term of the corresponding main level, if present, followed by ‘l’ (local). If the character “p” for “phrase” is used also, the character “l” is preceded by “p”.
- The search terms of the copy indexes start with the search term of the corresponding main level, if present, followed by ‘c’ (copy). If the character “p” for “phrase” is used also, the character “c” is preceded by “p”.
- The search keys for special general indexes for a CBS of CBS MARC 21 fields that are a 'variant' of existing indexes use the search attributes of those indexes, followed by 'x'.
Search attributes
CBS MARC 21 uses the search attributes, as used for WorldCat, for word indexes only, unless a search attribute is defined for a phrase index only; when using the phrase version of an index the number for the search attribute is increased by 10000.
E.g. search attribute 4 is related to the title index. In a CBS MARC 21 database the word index uses search attribute 4. But for the corresponding phrase version we use search attribute 10004.
To be as complete as possible, we will also use the Z39.50 search attributes. That is why from time to time, you will find two items for an index in the attribute column.
For the search attributes of the CBS MARC 21 indexes, the following ranges have been determined:
- 00001-09999
For word indexes for CBS MARC 21 fields that are based on equivalent WorldCat attributes.
Remark: For the CBS MARC 21 system, several search attributes are hard-coded. E.g. the range 90XX is reserved for special purposes.
- 10001-19999
For phrase versions of indexes for CBS MARC 21 fields, the search attribute of the word version is used increased by 10,000.
- 20001-20999
- For indexes for CBS MARC 21 fields, of which the corresponding WorldCat search attribute is absent on the OCLC Index Documentation webpages.
- This applies to all levels: main, local as well as copy level fields.
- Special general indexes for a CBS of CBS MARC 21 fields on main level.
- 21001-22999
- Indexes for CBS MARC 21 local level fields (LBD).
- Special general indexes for a CBS of CBS MARC 21 fields on local level.
- 23001-24999
- Indexes for CBS MARC 21 copy level fields (LHR), as far as not indexed in the ranges 00001-09999 and 10001-19999.
- Special general indexes for a CBS of CBS MARC 21 fields on copy level.
- 30001-31999
- Indexes for non-CBS 21 MARC fields for all levels, i.e. main, local and copy level.
- These fields can be split into the categories system fields and non-system fields.
[1] November 2020: The software and configuration need to be adapted to make it possible to use the colon (:) to search word indexes and the equal sign (=) to search phrase indexes. When this is realized, the “p” will not be necessary anymore.