GeographicPlaces
From oaibp
Main Page >> Shareable Metadata
Note: Summary of best practices added. Jenn Riley 10/12/05
Contents |
[edit] Recording Geographic Places in OAI Records
[edit] Summary of Best Practices
- Choose geographic place values from relevant controlled vocabularies consistently and explicitly.
[edit] Choice of Geographic Place Names and Locations
Specificity
In choosing a geographic place name, the most specific place name or location available should be applied. Use of a controlled vocabulary can assist in providing disambiguated geographic locations. When a controlled vocabulary is not in use, or does not require disambiguation it is recommended that sufficient context be added to the location to differentiate it from other places with the same name.
For locations within the United States and Canada, provision of the name of the state or province level is sufficient. When an existing standard is not in place for providing this information, it is suggested that usage of a two character postal code indicating state or province be used for disambiguation.
Outside of the United States and Canada the country, and when necessary the next level of political subdivision, should be provided.
It is also recommended to apply as many geographic place name terms as necessary to describe the item in question.
Controlled Vocabularies
The best approach for supplying geographic place names or location information is to consistently and explicitly use relevant controlled vocabularies. The controlled vocabularies chosen (more than one may be used if appropriate) should be relevant to the resource described and frequently used within the community to which the materials being described would hold the most interest. Depending on the materials being described, vocabularies such as the Thesaurus of Geographic Names (TGN), the U.S. Geologic Survey's Geographic Names Information System (GNIS), or the Alexandria Digital Library (ADL) Gazetteer may be useful as controlled vocabularies for geographic location information.
It is also worth mentioning that the Library of Congress Subject Headings (LCSH) (also available online as part of the Library of Congress Authorities), contains geographic place name headings. When using a controlled vocabulary like LCSH, which mixes topical and geographic location terms in a single vocabulary, you should map terms that are unambiguously geographic in nature (e.g., Calgary (Alta.)) to a geographic place or location field in your metadata format if one exists, while mapping explicitly topical terms that include geographic subdivisions (e.g., Central business districts--Alberta--Calgary) to a subject or topical field in your metadata format. If possible you should also map the geographic subdivisions to the geographic place or location field in your metadata format if one exists. This allows service providers to take better advantage of the geographic names in their design of the search (or other) service.
[edit] Format and Encoding of Geographic Place Names and Locations
When using a metadata schema that allows repeating element, each distinct geographic place or set of geographic coordinates should be placed in its own element. For example, if you are describing a text which discusses both Santa Fe, New Mexico and Boise, Idaho these two distinct locations would be placed in two separate elements, rather than being packed into a single element.
Hierarchies
When encoding hierarchical data, it is recommended to encode the individual components of the hierarchy, if your metadata schema supports this.
Example: MODS is a schema that allows the encoding of hierarchies in geographic places:
<subject>
<hierarchicalGeographic>
<country>United States</country>
<state>Kansas</state>
<county>Butler</county>
<city>Augusta</city>
</hierarchicalGeographic>
</subject>
When mapping hierarchical geographic elements into a metadata schema that does not support the encoding of hierarchies, concatenate the components of the hierarchy into a single geographic element.
Example: The above example mapped into Dublin Core:
<coverage>United States--Kansas--Butler--Augusta</coverage>
Coordinates
Several metadata standards allow the encoding of geographic coordinates. The Dublin Core Metadata Initiative (DCMI) Point Encoding Scheme provides a alternate strategy for encoding a geographic location as "a point in space using its geographic coordinates," as does the DCMI Box Encoding Scheme for encoding a geographic location as "a region in space using its geographic limits." The Metadata Object Description Schema (MODS) provides elements for the encoding of cartographic data including coordinates, scale and projection.
[edit] Best
Ideally, the metadata format in which you expose your OAI records should allow the unambiguous specification of the vocabulary used as the source for each geographic place name applied to those records.
Example: MODS is an example of a metadata schema with an attribute (authority) that supports unambiguous specification of controlled vocabularies used, encoding of geographic hierarchies and encoding of geographic coordinates:
<subject authority="lcsh">
<geographic>Santa Cruz County (Calif.)</geographic>
</subject>
<subject authority="tgn">
<hierarchicalGeographic>
<continent>North and Central America</continent>
<country>Canada</country>
<province>Quebec</province>
<area>Matapedia, Lac</area>
</hierarchicalGeographic>
</subject>
<subject authority="gnis">
<cartographics>
<coordinates>36 58 27 N, 122 01 47 W</coordinates>
</cartographics>
</subject>
Qualified Dublin Core also allows the specification of controlled vocabularies used:
<dc:subject xsi:type="dcterm:LCSH">Central business districts--Alberta--Calgary</dc:subject>
<dcterm:spatial xsi:type="dcterm:LCSH">Calgary (Alta.)</dcterm:spatial>
<dcterm:spatial xsi:type="dcterm:LCSH">Alberta</dcterm:spatial>
[edit] Good
If limited to simple Dublin Core (oai_dc), or another metadata schema that does not support the explicit specification of the controlled vocabulary source(s) you used to apply each geographic place name, the next best option is to apply terms from a single standard controlled vocabulary consistently within an OAI set or an entire repository, and specify the vocabulary within the set description [[[SetPractices]]] or Identify response. See the section on providing [supplemental information] to service providers.
Examples: Simple Dublin Core (oai_dc) is an example of a metadata schema that does not support specification of controlled vocabularies used:
<dc:coverage>Santa Cruz County (Calif.)</dc:coverage>
[Note in set description or Identify response that LCSH is your controlled vocabulary for geographic place names in the <coverage> field.]
<dc:coverage>North and Central America--Canada--Quebec--Matapedia, Lac</dc:coverage>
[Note in set description or Identify response that TGN is your controlled vocabulary for geographic place names in the <coverage> field.]
<dc:coverage>36 58 27 N, 122 01 47 W</dc:coverage>
[Note in set description or Identify response that GNIS is your controlled vocabulary for geographic places in the <coverage> field.]
[edit] Adequate
At minimum, a metadata provider should use a local vocabulary or list to apply geographic place names in a consistent fashion across all records in an OAI set or across a repository. Geographic names should be placed in the appropriate geographic field where one exists.
In addition to the guidelines presented in this section, the RLG Descriptive Metadata Guidelines for RLG Cultural Materials http://www.rlg.org/en/pdfs/RLG_desc_metadata.pdf contains advice on the use of geographic place names in metadata records.
