Machine-Readable Bibliographic Information Committee

MARBI met twice at ALA Midwinter 2011 in San Diego to consider four discussion papers and one proposal. Full texts of the papers are available at:

A summary of the discussions follows:

Proposal 2011-01 dealt with a request from OLAC for a way to cleanly code for the original language(s) associated with moving images. There is currently no way to distinguish between the original language and intermediate languages. After discussion, the proposal was refined to redefine the $h subfield for use only when encoding the language of the originalís primary content. New subfields were also defined: $k to cover intermediate languages, $m for the langauge of accompanying material such as notes on containers and the like, and $m (?) for the language of a libretto. The revised MARC documentation will include a note saying that subfield $h is not required if the language associated with the resource is not a translation.

Discussion Paper 2011-DP03 concerned identification in bibliographic, authority and holdings records of the FRBR Group 1 entities (work, expression, manifestation, and item). This would facilitate experimentation with FRBR applications, such as FRBR'ized databases, and assist with validation, if and when FRBR compliance becomes a validation criterion. The applications do not yet exist, so this field is being defined for future use. The 883 was suggested as the field for this information, which could take the form of either a term or a code. Both the paper and the discussion acknowledged several difficulties with identifying Group 1 entities. Records currently contain information relating to more than one FRBR entity, for example, a bibliographic record may contain details that properly belong to all four levels. Moreover, the distinction between the different FRBR entities can be quite subtle, and catalogers may legitimately differ on which FRBR entity is represented by a particular resource. Although the field is intended to help catalogers interpret a record, it may just add to their confusion. To make the field useful, there need to be guidelines and definitions governing its use. It should also be defined for use only when the record represents only a single FRBR level (this would rule out its use for unique objects), and be made optional, so catalogers need not agonize over which FRBR level to assign. There was some discussion of how current cataloging practices do not jibe with the FRBR model: for example, it is unclear whether language belongs to the work level or the expression level. There is a case to be made for creating an expression-level reocrd for the work in its original language, as well records for subsequent translations. If this distinction was clearly made on authority records, it would facilitate ongoing updates; for example, vendors could be instructed to send updates to work-level records, but not to expressions, if the insitution does not own a copy of the translated version (there is no point in cross-references from a given language if you don't own a copy of the work in that language). The number of issues raised by the paper made it unclear how to proceed; several participants suggested that the analysis of RDA test records might provide data which could guide further development of the idea.

Discussion Paper 2011-DP04 dealt with the creation of new MARC subfields for carrier attributes defined in RDA. Most of these carrier attributes are already being recorded in MARC records, but they are not defined for separate subfields; for example, base material, applied material, and production method are all recorded in subfield 300$b. The paper suggests that new subfields be added to the 340 field to record the carrier attributes listed in the paper. To facilitate this, the definition of the field would be broadened beyond its current scope, which is to describe items that require technical equipment for their use or items with special conservation or storage needs. Some institutions are already using the 340 field for more granular recording of carrier attributes that do not pertain to technical equipment or special needs, so this would not be too much of a stretch. Some participants questioned whether this approach limits the extensibility of new carriers and formats, as we would soon run out of subfields. But it was pointed out that the carrier attributes are not characteristics of formats, but methods of sensory transmission; and barring some great evolutionary leap or mutation, these are fairly stable and limited in number. Nonetheless, technical metadata is dodgy because so volatile, and the MARC community might eventually opt to pull these elements out again. There was general agreeement that the issue was worth pursuing; when the paper comes back before the committee, specific examples from various formats would be extremely useful.

Discussion Paper 2011-DP01 dealt with new data elements needed to accommodate RDA production, publication, distribution and manufacture statements, and copyright notices. RDA defines separate elements for production, publication, distribution, and manufacture statements, including place, name, and date. This level of granularity is currently not supported in MARC. The issue is complicated by the fact that MARC will have to accommodate legacy data, and also permit catalogers to continue recording undifferentiated data in situations where the information on the resource is unclear, or where the distinction between manufacture and publication is not meaningful, as in pre-modern publications. The paper offered several different options for distinguishing between the different types of production information: continue to use the 260 field, but use the second indicator to identify the type of function (option 1); continue to use the 260 field, but add a separate subfield for each function (option 2); define new fields for each function (option 3). All solutions would have to build in a method of indicating that no distinction is being made between the different functions. There was little support for option 2 (too messy), and an interest in combining options 1 and 3. This would entail continued use of the 260 field as currently defined, and creation of one new field with an indicator specifying the function. Copyright notice statements will have to be dealt with in some other fashion, probably by defining a new 2XX field. The issues will be revisited at Annual.

Discussion Paper 2011-DP02 dealt with additional elements needed in the MARC bibliographic and authority formats to record RDA data elements such as language of expression, language associated with a person or corporate body, associated institution, fuller form of name, and type of jurisdiction. These elements are required in headings, if needed to disambiguate a heading. It would be up to the institution to decide whether to also include the information in the newly-created fields, whether or not they were also used in the headings.

The paper considered two options for language of expression: using the 041 on both the bibliographic record and the authority record (with fewer subfields, since the 041 in the bibliographic record covers manifestatons). The other option would be to broaden the definition of the 377 field (Associated Language) to include language of expression as well as language associated with persons or corporate bodies. There was no consensus on which approach was preferable.

Associated institution information, which is a core element for headings for conferences, etc. if the institutionís name provides better identification than the local place name, could be accommodated in the 373 field. This field is currently defined for Affiliation, and limited to groups with which a person is affiliated; a new subfield for institution associated with corporate body could be added.

The fuller form of name is currently required only if needed to break a conflict. Recording it in a non-heading field, if known, would provide catalogers with useful information, and a means of resolving future conflicts. A new field, the 378 field, was proposed for this information, which can pertain to either the surname or the given name of a person.

Type of jurisdiction is another element used to break a conflict, e.g. Cork (Ireland : County), as opposed to Cork (Ireland) (the heading for the city). A new field, 334, was suggested for this information. It could be very useful for identifying geographic areas that are non-governmental, such as rivers and mountainsópreferably using terms from a controlled list, such as the Board of Geographic Names list.

Participants also noted that there are other conflict resolvers which are not currently coded, such as form of work. FRAD also does not include some attributes which might be relevant, such as language of family.

Report by Elizabeth O'Keefe
ARLIS/NA representative to the MARC Advisory Committee

... go to more ALA reports ...