A Historian's Perspective on Georeferencing and Interoperability

Ruth Mostern, UC Merced

• Digital gazetteers
  • a structured dictionary of named places
  • can be interegrated with network-accessible svcs to respond to queries and retrieve geometries
  • valuable information resources in their own right
  • and provide thte basis for place-based search, display and integration

• another perspective on place
  • Yi-Fu tuan, "Place: An Experiential Perspective" (1975)
    • place as a center of meaning, constructed by experience
    • is it really true that gazetteers can't encompass all place information

• Chinese sacred geography
  • emei mountain in the foguang encyclopedia (secondary reference source):
    • 5 lines of text in the entry that presents: 25 places named in a single entry, in addition to the headword; an additional 210 places on the mountain referenced by category (temples, peaks, etc.)
    • how it describes it as a center of meaning
      • location in political space
      • alternate names
      • neighboring topographical features
      • constituent features (peaks)
      • status in corporate group of Four Great Sacred Mountains
      • significant named built features on the mountain
      • other features referenced, e.g., 70 temples, 40 grottoes
  • descriptive gazetteers as historical sources (a 13th century example), giving the history of a praefecture
    • example: 2 new counties created in the praefecture in a particular year, addressed by name

• mapping the foguang encyclopedia
  • Willard McCarty? 2005: McCarty? suggests that “the twin computational requirements of complete explicitness and absolute consistency” opens up a space for the scholar “to refine an inevitable mismatch between a representation and reality” in a way that other modes of production do not.

• implications for gazetteer design
  • historical gazetters need to support content that is:
    • multilingual (language, script, translation issues)
    • temporally intricate (not just start and end date, but individual names have dates, individual footprints are temporally limited)
    • spatially and temporally ambiguous
    • richly attributed in textual sources
    • reflective of indigenous ideas of space
    • of diverse scale

• perspectives on interoperability
  • gaz as infrastructure
    • integrating with, or using in, other types of systems and resources
    • e.g., ecai work in the last year or so - linking who, what, where and when resources as a way into specific information about libraries' websites and other sources
  • gaz and other knoledge organization systems ("time-ateers")
    • an event (a state of being) is described by:
      • title
      • alternative names
      • start and end dates
      • attested dates
      • location
      • relationships
      • bibliographic information
      • creator
  • single socially authored gazs
    • a gazetteer environment that allows people to work
    • multiple validated contributors with or without editor
    • associated discussion site
    • edited wiki (editorial control?)
    • free-for-all wiki
  • multiple fedeerated gazs
    • rich with relationships
    • most historical gazs will emerge as small, handmade, specialist artifacts - let's put them together!
    • but ....
      • incompatible typing schemes
      • unmanageable ambiguity
        • problem even in a single gazetteer
        • what happens when the domain is multiple gazetteers that treat "the same place" but were altered and organized by separate authors
      • conflicting authorities
        • were the mongols a great civilizing influence or a bunch of barbarians?
        • how to you "federate" a palestinian and an israeli gazetteer?
      • challenge of achieving buy-in
        • what's the academic and personal reward structure
        • it's not just about making it easy

Comments

• RF@NGA:
  • NGA has many other databases that we do not think of as gazetteers, but in some way (this context) they really are
  • relating places and events about places and keeping that in some kind of structured knowledge organizational system
  • given that it relates to palce, we can apply the gazetteer paradigm

• ??
  • context of inclusions (socially constructed gazetteers)
  • emergency services: difficulty of separating the collective knowledge required wihtout compromising security and privacy concerns
  • buy in is an issue here

• ??
  • "unmangeable" ambiguity ... an objection
  • ambiguity is complicated, but it's not necessarily unmangeable
  • you have to research the ambiguity to figure out if you can mange it
  • it's not necessarily the case that we can't manage it
  • supposedly intractable situations may not be ... look first

• Bruce
  • place ambiguity - you simply don't know whether it refers to one or another mountain
  • those kinds of ambiguity are not fully tractable

• ??
  • the "time-ateer" -- those types of ateers may be a way to get to a concept of the "minimal" gazetteer
  • the "knowledge organization system"
    • a gaz is one type of a more generic idea that is modeling and structuring information in some domain
    • is this a genre?

• ??
  • gazetteers are culturally grounded, and given if you can lie with maps, can you lie with gazetteers?
  • what are the consequences for modeling
  • how do you put an imperical measure on the cultural distortion
  • in a historical context - fidelity to original sources and rich attribution (footnoting)
    • you must indicate where you got the information and why you interpreted it the way you did
  • a variation ont he trust issue, metadata is a touchstone issue here
  • wider issue for later discussion: do we think the gazs we currently have have adequate metadata
    • in many cases, they are argueably not rich enough in this regard

• ??
  • disconnect between various disciplines that could contribute
    • particularly with historical cultural gazetteers - it gets more qualitative and relationship-centric
    • ontology is more attractive in this domain
    • reasoning systems to support ontologies - but they seem to be generally confined to "practical" domains (context here is AI/CompSci)
    • is there any such Knowledge Organization support technology that's actually been applied to a gazetteer
      • ??: Natl Counterterrorism Center - an ontology that doesn't know enough about the linkages
        • rich attribution for new pieces of information is critical
        • every time a new linkage is created, they have to explain it in a rich way
      • there is an old one
        • police application systems that are about who did what to whom when where - solving crimes: events, places and relationships
        • e.g., Cop Link

• RM
  • thinking rigorously about relationships
  • we know how to do attributes of objects
  • how do we attribute the relationships so that we make them rich, interesting, consisten and useful

• a number of national mapping agencies are developing ontologies
  • being very explicit using tools comeing out of ai research (e.g., OWL)
  • they are not deliberately addressing teh gazetteere, but they are building up the ontologies of types and relationships for the database to which the gazetteers refer
    • there's a lot of potential synchrony here

• ??@redlandsinstitute
  • exploring these issues,w ith a focus on relationships between objects
  • tortoise habitat management project - data includes images and text documents, as well as gis data
  • have been using a gazetteer modeled after adl - but user community needs more than the gaz provides
  • what's relationship betwen ontology and the gazetteer
    • can in theory automatically generate and maintain the gazetteer out of the more comprehensive ontology

• ??
  • are gazetteers infrastructure?
    • functional decomposition fo the existing protocols to support a richer suite of services
    • is it a matter of putting an individual gazetteer into an atlas system - all custom-built stuff - a differnet question from Greg's use scenarios
  • how do we think locally and build globally?