Z39.50 Configuration Guide for OCLC Z39.50 Cataloging

This Configuration Guide contains technical specifications for connecting a Z39.50 client to the OCLC Z39.50 server to use the OCLC Z39.50 Cataloging service.

The OCLC server implements Z39.50-1995 (version 3 of the protocol).

Please configure your client and test it on our test server. You need a test authorization to access the test database. To obtain a test authorization or to ask questions, contact OCLC Support.

For information describing MARC record structure, character sets, and exchange media formatting, see OCLC-MARC Records .

Internet Addresses and Hours of Availability

Database Names

OLUCWorldCat
OCLCAuthoritiesLC
OCLCInstitutionRecords

Production Server

Hours

Test Server

Note: The Test Server may not be available on Monday mornings. Otherwise, every effort is made to keep it available when the Production Server is available but we cannot guarantee this availability.

Z39.50 Services Supported

The OCLC Z39.50 server supports the following services:

• Initialization Service
• Search Service
• Present Service
• Scan Service (From Z39.50-1995)
• Extended Services (From Z39.50-1995)

Initialization Service

• Protocol Version - Z39.50-1992 (Version 2)
• Options Supported
• • Search
  • Present
  • Scan (From Version 3)
  • Extended Services (From Version 3)
• Preferred Message Size - OCLC will accept the client's proposal for Preferred Message Size up to a limit of 105000 bytes.
• Maximum Record Size - OCLC will accept the client's proposal for Maximum Record Size up to a limit of 105000 bytes.
• ID Authentication - OCLC supports ID/authentication as prescribed by the standard. We recognize the userId and password fields; the groupId field is ignored. The service patron password is the password that must be used.
• Character set negotiation - OCLC supports both MARC-8 and UTF-8 (OID 1003.15.1 and 1003.15.3) character sets. The default is MARC-8. Character set is specified in the otherInfo field.
  Note: When the MARC-8 default is retained, characters that cannot be converted to MARC-8 are displayed as hexadecimal numeric character references per MARBI proposal 2006-09 ( view example).
• userInformationField - On the Init Response message, we return the OCLC_Information Record in the userInformation field. The OID for this is 1.2.840.10003.10.1000.17.1

  OCLC-UserInformation ::= SEQUENCE {
     motd [1] IMPLICIT VisibleString,
     dblist SEQUENCE OF DBName,
     failReason [3] IMPLICIT SEQUENCE {
        diagnosticSetId OBJECT IDENTIFIER OPTIONAL,
        code [1] IMPLICIT INTEGER,
        text [2] IMPLICIT VisibleString OPTIONAL
     } OPTIONAL
  }
  
  DBName ::= [2] IMPLICIT VisibleString

• otherInformationField (Request) — On the Init Request message, OCLC supports 1.2.840.10003.15.1   and 1.2.840.10003.15.3 character set negotiations for UTF-8. Language set negotiations can be included, but are ignored.
  
  otherInfo {
        externallyDefinedInfo {
           OID: 1 2 840 10003 15 3
           externallyDefinedInfo choice
           {
              proposal {
                proposedCharSets {
                  iso10646 {
                    encodingLevel OID: 1 0 10646 1 0 8
                  }
                }
              }
           }
         }
     }
  }
  
• otherInformation Field (Response) — On the Init Response message, we return the externallyDefinedInfo field for a successful character set negotiation. If a language set negotiation was included, it was ignored and NOT returned.

  externallyDefinedInfo{
     OID: 1 2 840 10003 15 3
     externallyDefinedInfo choice
     {
       response {
         selectedCharSets {
           iso 10646 {
              collections OID: 1 0 10646 1 3 1 3
              encodingLevel OID: 1 0 10646 1 0 8
           }
         }
       }
     }
  }

Search Service

• Database-Names - OLUCWorldCat or OCLCAuthoritiesLC (case-insensitive)
• Result Set Name - The server maintains any number of named results sets during the Z-association.
• Element Set Names - Please see the section on Element Set Names Available.
• Query - Type-1 & Type-101 are supported.
• Attribute Set ID - Bib-1 only is supported.
• Operand - The Result Set ID (i.e., "default") can be used in the Operand (for qualifying searches).
• Operators Supported
• • AND
  • OR
  • AND-NOT

Present Service

• Additional-ranges - This feature is not implemented by our server.
• Result-set-id - The server attempts to find the named result set. If the result set is not found, the server returns an error 30. If the Result-set-id is not supplied, the server gets records from the last result set created.

All other fields behave exactly as prescribed by the standard.

Scan Service

• Database-Names - OLUCWorldCat or OCLCAuthoritiesLC (case-insensitive)
• Step-size - The OCLC server always uses a step-size of 0.

All other fields behave exactly as prescribed by the standard.

Extended Services

See Z39.50 Database Update Implementation for OCLC Z39.50 Cataloging: Extended Service Definition.

Record Syntaxes Supported

USMARC

Element Set Names Available

*Additional element set names are supported for non-Latin data.

**Default holdings. When you choose default holdings, the default is determined as follows:

Recommended Searching Combinations for the OLUC WorldCat Database

The following guidelines help you construct sample author, title, subject, standard number, and music number searches for bibliographic records with local Z39.50 clients. Remember that OCLC Z39.50 Cataloging supports only the Use Attributes Supported listed in this guide and only the default option for several attributes.

Author search (word)

The following sample uses a word search for the author Scott Adams.

Author search (word list)

The following sample uses a word-list search for the author Scott Adams.

Author search (phrase)

The following sample uses a phrase search for the author Scott Adams. Because right truncation is used, The search retrieves author entries for adams, scott with and without dates.

Title search (phrase)

The following sample uses a phrase search for the title Angela's Ashes..

Subject search (phrase)

The following sample uses a phrase search for the Library of Congress subject heading computers and women.

Standard number search (ISBN)

The following sample uses a word search for the ISBN 0679429220 (assigned to Midnight in the Garden of Good and Evil).

Music number search

The following sample uses a word search for the music number D2-24935 (assigned to Yodeling the Classics).

Use Attributes Supported

Refer to Searching Tips for OCLC Z39.50 Cataloging for a complete list of attributes supported.

Relation Attributes Supported

The OCLC server supports the following relation attribute values for cataloging purposes. The default value is equal.

Position Attributes Supported

The OCLC server supports the following position attributes. If no attribute value is sent, our implementation performs searches as if the value were any.

Structure Attributes Supported

For any Use attribute where we specify that word is a valid structure attribute, the values of word list (6) and ordered word list (7) are also valid. We implement word list as "AND" within a field. We implement ordered word list as "AND" with a proximity of 1 with order being important. Date (5) is implemented much like word with the most useful difference being it allows for ranging of dates (i.e. 1987-1993 will only hit on records between 1987 and 1993 inclusive.)

Truncation Attributes Supported

The OCLC server supports the following truncation attributes. The default value is attribute 104.

Completeness Attributes Supported

The OCLC server supports the following completeness attributes. If no attribute value is sent, our implementation performs searches as if the value is incomplete subfield.

See also Z39.50 Database Update Implementation for OCLC Z39.50 Cataloging: Extended Service Definition

Returning Non-Latin Data in Response Records

To select an option for display of non-Latin data Z39.50 clients may use either of these requests:

• Init (groupID field)
• Search/Present (ElementSetName)

The non-Latin values in groupID and ElementSetName are not case sensitive. No white space is allowed around the equal sign in the Init strings. Use of ElementSetName in Search/Present requests is preferred because it is more compliant with the Z39.50 standard.

If neither type of option is specified, the default is to include non-Latin data in 880 fields (Option 1 below). If a valid non-Latin option is found in both the Init and in a Search/Present requests, the OCLC Z39.50 server uses the option in the Init request, overriding any option found in a Search/Present request. Note: Any unrecognized value found in groupID or in ElementSetName is ignored.

Authentication. The OCLC Z39.50 server does not currently process the groupID field for authentication. The server will recognize the optional Init request substrings anywhere in the authentication groupID field, so a non-Latin specification does not need to be the only data in the groupID field.

Options: