DLF Aquifer MARCXML to MODS stylesheet
The purpose of the MARC-MODS conversion information provided by the Metadata Working Group is to assist those wishing to convert MARC records to MODS records suitable for use with OAI or other sharing implementations. Being congruent with the Guidelines, these records are designed for this purpose of sharing with other institutions, and thus may not preserve metadata necessary for use in other contexts.
We began from the MARC-to-MODS stylesheet (Version 3.2 revision 1.15) provided by the Library of Congress (mappings, stylesheet), customizing the mappings and stylesheet to match our guidelines. Important to note is that the stylesheet does not attempt to populate the MODS record with every element in our guidelines. Specifically, attempts were not made to add mappings to create elements we designated as "optional."
Many of the divergences from the LOC documents revolve around metadata describing related items or other versions. In many cases, the MODS paradigm as expressed in our guidelines did not allow for a clean crosswalk, and the conversion was left out to avoid inconsistent performance. In other cases, the information was deemed not necessary or applicable for the OAI environment or the object types likely to be shared via OAI.
We realized that we cannot attempt to create one stylesheet, or even a series of stylesheets, that would appropriately handle every possible variation in the use of MARC. Initially we have created a single stylesheet designed to transform records that follow the AACR2 content standard, with this exception:
Where a digital object is treated in the MARC record as a reproduction of an analog (physical) object, the stylesheet is designed for records that follow the tagging approach outlined in Library of Congress Rule Interpretations section 1.11A ; stated simply, this approach gives information about the original in many areas of the record, such as the title and statement of responsibility, publication area including dates, and physical description, and gives details that are particular to the digital reproduction (such as the digital publisher) in a note (533).
This stylesheet will also be appropriate for use with records that have followed a "single record" approach as outlined in LC's "Draft Interim Guidelines for Cataloging Electronic Resources".
If your records have followed other approaches, you may need to modify the stylesheet in order to appropriately map data. Please consult a member of the Metadata Working Group if you have specific questions.
Mappings that differ from LOC mappings:
007 and 008 codes mapped to a term in <genre>
Now use authority attribute "marcgt" to conform with MARC Code Lists (http://www.loc.gov/marc/relators/relasour.html#rela655 ) for MARC Genre Terms.
007 position 1 (Specific Material Designation)
No longer mapping value r (Remote) to <physicalDescription><form>; since all Aquifer resources are remote, this is unnecessary.
008 pos.07-10 (Date1)
This data is mapped to <dateIssued>; our stylesheet adds the attribute keydate="yes". This date is favored over the information elsewhere in the MARC record as it follows an encoding standard.
024 (Other Standard Identifier)
We have added mapping of identifiers with second indicator 7 (source specified in subfield 2) to <identifier> and subfield 2 contents to the type attribute.
210 (Abbreviated Title)
We have modified the stylesheet to map subfield 2 (source) to the authority attribute of <titleInfo>.
240 (Uniform Title)
We have included mapping of subfield s (Version) in sequence, to <title><titleInfo type="Uniform">
242 (Translation of title)
We have added mapping for initial articles to <title><nonSort> based on 2nd indicator value.
250 (Edition Statement)
The current mapping is for subfield a only; we have changed this to add contents of subfield b to subfield a in the <originInfo><edition> element.
300 (Physical Description)
We have added mapping of subfield f (type of unit)
Excluded mapping of this field to <note> for serials, but retained it for other formats. Important bibliographic citations in monograph records, especially for older books, may be useful to researchers. Serial records are more likely to have lists of indexes which are less useful.
533 (Reproduction) subfield f (Series statement of reproduction)
1. Change mapping from <note type="reproduction" to <note displayLabel="reproduction">.
2. In addition to current mapping of all subfields to a <note> element, we add mapping of subfield f to <relatedItem type="series"><titleInfo><title>.
245 subfield c (Statement of Responsibility)
506 (Use Restrictions)
533 (Reproduction) subfield f (Series statement of reproduction)
538 (System details)
Due to a more conservative approach on allowed types, we have changed these mappings from <note type="xxx"> to <note displayLabel="xxx">
561 (Ownership/Custodial History)
Mapped to <note>; added attribute displayLabel="ownership".
581 (Publications about Described Materials)
Mapped to <note>; added attribute displayLabel="publications".
We include mapping of second indicator codes to the authority attribute.
662 (Subject Added Entry - Hierarchical Place Name)
Added this tag, mapped to <subject><hierarchicalGeographic>
752 (Hierarchical Place Name)
Added mapping of subfields f, g and h to <hierarchicalGeographic> subelements <citySection>, <region>, and <extraterrestrialArea>. See also note about 752 under "Mappings requiring your attention".
762 (Subseries entry)
This is currently mapped as <relatedItem type="series">. However, it describes a sub-series of the item described by the record (a series or serial). This situation fits better with type="constituent".
765 (Original language), 767 (Translation)
Currently mapped to <relateditem> without a type attribute; we add type="otherVersion".
773 (Host item)
When leader 07 (BLvl) value is a (monographic component part) or b (serial component part)* In addition to the mapping to subfields of <relatedItem>, we also map 773 subfield d to <originInfo><publisher>. Note that subfield d contains place and date of publication in addition to publisher. If your institution already has publisher information in tag 260 for "In" analytics, you may wish to modify the stylesheet to omit the additional mapping of 773 to <originInfo><publisher>.
856 subfield u (Electronic location/access, URI)
When a URL which is mapped to <identifier> contains an ARK identifier, we assign type="ark" to the <identifier> element; this is similar to the treatment for DOI and Handle.
856 (Electronic location)
Desiring to create logic to assign URLs a "primary display" usage attribute, we needed to make some assumptions about records that incorporated multiple URLs under the 856. If there is no indication of a primary display, and multiple URLs with the same indicator are present, we assume the first URL is the primary display. We interpret an 856 with the second indicator 1 ("related resource"), if the sole 856, as pointing to the main display as well. In other situations, order of preference is an 856 with indicator 0, then 1, then blank.
Mappings requiring your attention:
533 (Reproduction) and 538 (System Details)
These fields may be contain formal, standardized notes if the record has been used for the DLF Registry of Digital Masters. We have left the mapping to <note> in the stylesheet since these fields may have other uses. An institution may choose to modify the stylesheet to omit these fields if it believes the standard notes would not be useful to Aquifer.
752 (Hierarchical Place Name):
This tag is mapped to <subject><hierarchicalGeographic> because, until tag 662 Subject Added Entry - Hierarchical Place Name was approved, this tag could be used for subjects. If the cataloging practice for a particular group of records has been to use this field to specify some other geographic aspect of the item (such as place of creation), the stylesheet may need to be modified to remove this mapping.
856 subfield q (Electronic Format Type):
Although not explicitly required by MARC or AACR2, we are assuming that data in 856 subfield q (Electronic format type) is a MIME Media Type and have mapped it to <physicalDescription><internetMediaType>.
MARC Tags left unmapped due to not being useful in providing metadata in an aggregated environment:
510, 530+, 777, 786.
+530 (Additional Physical Form Available Note): If your cataloging standards include useful information here, you will need to edit the stylesheet to include a mapping to <relatedItem type="otherFormat><note>.
MARC tags left unmapped due to not providing enough information to provide a required "type" attribute:
787, 856 _2.
Appendix: More information on MARC tags left unmapped
530 (Additional Physical Form Available Note):
Not likely to be useful in the context of aggregated metadata. If your cataloging standards include useful information here, you will need to edit the stylesheet to include a mapping to <relatedItem type="otherFormat><note>.
534 (Original version)
Use of this tag appears to be uncommon and deprecated. Our stylesheet assumes that if a record for an "analog" object treats the digital object as a reproduction, it does so by noting it in 533.
583 (Action note)
This tag is used to record information about the institution's processing with respect to the object, not the object's content.
777 (Issued with)
This level of detail is beyond the scope of what Aquifer is likely to provide.
786 (Data Source)
This describes the source of data for some geospatial applications, and is beyond the scope of the type of information Aquifer is likely to provide.
787 (Nonspecific relationship)
Being unspecific by nature, this does not meet our requirements to provide a type attribute; nor is it likely that this data would be important enough to include in the Aquifer record.
856 _2 (Electronic location/Access, related item).
Aquifer MODS guidelines require a type attribute for <relatedItem> and there is no existing "type" that fits this element.