Records in the AMICUS database do not conform to one specific cataloguing standard. The database contains records from distribution sources such as the Library of Congress, Library and Archives Canada, CONSER, records created on-line by participating member libraries, including the print accessions received and input by the Union Catalogue of Library and Archives Canada, and records contributed by libraries participating in the MARA program. These records exhibit a variety of cataloguing practices, but for the most part are consistent with the Anglo-American Cataloguing Rules (AACR), 1st and 2nd editions, and Règles de catalogage anglo-américaines (RCAA), 1ère and 2ième édition. (5),(6) Records reported via MARA must conform to AACR2 conventions, though records catalogued prior to the implementation of AACR2 are exempt from this criterion.
Institutions may submit MARA records catalogued in either English or French, regardless of the language of the publication being catalogued.
Contributing libraries may have used various romanization tables in their cataloguing. Canadiana, the national bibliography, uses the ALA-LC romanization tables for records catalogued in English and the ALA-LC romanization tables or the ISO transliteration tables for records catalogued in French. MARA libraries may conform to this usage, but it is not a requirement. (7),(8)
MARC or MARC-like records submitted as MARA records must be convertible to the MARC 21 Format for Bibliographic Data. The content designators-codes and conventions that identify the actual data within the record and that support data manipulation-are defined in this Format. The Format also identifies the authorized content requirements for all MARC elements (e.g., tags, indicators, subfields).
The key to retaining the MARC structure while reducing required specific coding is the "fill" character. Presence of a fill character indicates that the MARC format requires a value that the reporting library has not supplied: the fill character replaces a code that represents explicit data. The use of the fill character is limited to tags, indicators, subfield codes, and fixed tags only as indicated in the MARC 21 Format for Bibliographic Data. The fill character is represented in this guide by the "at sign" (i.e., "@").
Data required to support the distinction between bibliographic items have been defined as "mandatory" or "essential if applicable".
In submitting a MARA record, the reporting library must supply those data elements defined as "mandatory". Elements defined as "essential if applicable" must be present if they apply to the type of record being reported (e.g., monograph, continuing resource, sound recording) or if they are needed to fulfill bibliographic requirements (e.g., dates of publication for continuing resources).
If "mandatory" and "essential if applicable" elements are not supplied in a MARA record, or if a MARA record contains invalid content designation for these elements, then either the record will be rejected during loading, resulting in error logs that must be processed manually by the Union Catalogue, or the record will not be defined adequately enough to allow the AMICUS matching programs to work properly during loading. When the loading programs cannot correctly identify a record, the result is either a merger with a record for a different bibliographic item (i.e., a mismatch), or the creation of a duplicate record in the database.
For the minimal amount of data required in a full MARA record see "Appendix A: Required data elements".
The common standard used in transmitting a MARC record is the American Standard Code for Information Interchange (ASCII) character set, which includes upper case and lower case alphabetic characters. Bibliographic data elements in a MARA record should be reported in upper case and lower case. Subfield codes are written in lower case alphabetic characters or numerals. Values present in the leader are also in lower case.
The fill character is represented in ASCII by the "vertical bar" (i.e., "|") corresponding to hexadecimal "7C".
A MARC record consists of four sections: leader, record directory, control fields, and variable tags.
The leader is of fixed length and contains coded information relating to the processing of the record.
The record directory contains fixed-length entries for each tag appearing in the record. These entries, 12 characters each in length, define the tag and identify the length of the tag, and its starting position in the record. The record directory enables data elements in the record to be distinguished from each other, and to be machine manipulated.
Control fields are three-digit tags that take the form 00X (i.e., tags 001-009; tag 000 is reserved for the leader). These fields contain control numbers (e.g., tag 001) or may contain coded information in a fixed-length field (e.g., tag 008), which contains information regarding the item being described.
Variable fields are identified by three-digit tags, and carry bibliographic statements associated with the item being catalogued (e.g., author, title). Two numeric indicators are associated with each variable field. They either address the processing requirements of, or give further information about, the data in the field. Subfield codes within each tag (e.g., $a, $b) further identify the data elements, such that they may be distinguished and machine manipulated.
The Library of Congress' Web-based tutorial--Understanding MARC Bibliographic: Machine-Readable Cataloging--is an excellent source of information on the nature and structure of the MARC format. It can be accessed at: www.loc.gov/marc/umb/. (9)
To report their records via electronic file transfer, a library must have File Transfer Protocol, Transmission Control Protocol/Internet Protocol (FTP/TCPIP).
The following provides the set-up information for the MARA file transfer service.
The following sign-on parameters must be kept in lower case:
File naming conventions
Any combination of alphabetic or numeric characters, underscore, or period is valid. There are no upper case or lower case restrictions.
Short unique file names of 15 characters or fewer are preferred.
Batch numbering is recommended in order to distinguish one file from another (e.g., "MARA-batch1", "MARA-batch2", etc.). Other examples of meaningful and valid file names could be:
Test files require special attention and are generally the first files sent by a library when they join the MARA program. Test files must also be sent when a library changes their library automation system.
An underscore character (i.e., "_") in the first character position of a filename automatically identifies it as a test file. For example:
Retrospective files ("retro files") are files that contain the full content of a given library's database. Retrospective files are typically received in two situations: a) when a library that has never before reported to the National Union Catalogue sends the entire contents of its database to the National Union Catalogue for the first time; or b) when a library that has previously sent records to the National Union Catalogue wishes to "refresh" its holdings in AMICUS by first stripping its existing holdings out of AMICUS, and then reloading its entire database into AMICUS.
Underscore characters (i.e., "_") in the first two character positions of a filename automatically identify the file as a retro file. For example:
File Size Restrictions
The preferred file size is between 40,000 and 60,000 records, as the transfer of very large files requires monitoring by Library and Archives Canada's Information Technology Services branch.
File Transfer Mode
Files must be sent using the BINARY transfer mode to preserve the MARC format. Libraries that wish to send very larges files (e.g., a retrospective file or a large backlog) may do so only by special arrangement with the Union Catalogue.
Libraries that wish to report via another medium (i.e., not by FTP), must first consult with the Union Catalogue to determine whether their file will be readable once received.