[Ver: A3B2 8.07r/W] File: {UTPress}Cartographica/CARTO-40(4)-005.3d
Creator: / Date/Time: 8.11.2005/10:38am Page: 47/62
Social Perspectives on Semantic Interoperability:
Constraints on Geographical Knowledge from
a Data Perspective
Nadine Schuurman
Department of Geography / Simon Fraser University / Burnaby / BC / Canada
Abstract
Much attention has been paid by government agencies and GIS researchers to standardization of data and interoperability
of systems. Many of these efforts, however, have focused narrowly on technical hurdles while ignoring the social and
political contexts that influence interoperability decisions. This article illustrates how social factors influence
interoperability along three axes: classification, ontologies of data models, and government policy. Extant research
approaches to interoperability of GIS are discussed and their strengths and weaknesses assessed. The article begins with
definitions of what interoperability is, why it is important to academic users and policy makers, and its influence on
geographical knowledge in a digital age. Exploration of social influences, as an alternative analytical approach to
interoperability, begins with a discussion of the roles of classification and scale. The dangers of maintaining inflexible
ontologies associated with specific data models are illustrated as a technical limitation with profound social implications
for the construction of knowledge. Finally, policy at the multiple levels of governance with respect to interoperability is
explored as an infrastructural constraint – and a diminishing influence.
Keywords: data standardization, GIS, interoperability, semantic interoperability, spatial databases
Résumé
Les organismes gouvernementaux et les chercheurs dans le domaine des SIG ont porté une grande attention à la
normalisation des données et à l’interopérabilité des systèmes. Toutefois, la plupart de ces efforts ont été particulièrement
concentrés sur les difficultés techniques sans tenir compte des contextes sociaux et politiques qui influent sur les décisions
en matière d’interopérabilité. Dans l’article, on peut voir comment les facteurs sociaux influent sur trois aspects de
l’interopérabilité : la classification, les ontologies des modèles de données et les politiques gouvernementales. Les
démarches qui existent encore en matière d’interopérabilité des SIG sont passées en revue, avec leurs points forts et leurs
points faibles. Au début de l’article, on définit l’interopérabilité, son importance pour les utilisateurs en milieu universitaire
et les décideurs, et ses effets sur la connaissance géographique à l’ère numérique. L’exploration des influences sociales, en
tant qu’autre approche analytique en matière d’interopérabilité, commence par une discussion sur les rôles de la
classification et des échelles. Les dangers associés au maintien d’ontologies inflexibles liées à des modèles de données
précis représentent une limite technique qui a des conséquences sociales profondes sur l’approfondissement des
connaissances. De plus, l’auteur explore les politiques sur divers plans de gouvernance pour ce qui est de l’interopérabilité
en tant que contrainte dans l’infrastructure, et en tant que sphère d’influence qui perd de l’importance.
Mots clés: normalisation des données, SIG, interopérabilité, interopérabilité sémantique, bases de données spatiales
Data exist in a folded space and time, torqued by practices and problematics of local scientific communities that
produce them. (Bowker 2000)
cartographica (volume 40, issue 4)
47
[Ver: A3B2 8.07r/W] File: {UTPress}Cartographica/CARTO-40(4)-005.3d
Creator: / Date/Time: 8.11.2005/10:38am Page: 47/62
Nadine Schuurman
Introduction
GIS interoperability and semantic standardization are the
sorts of topics that halt conversations and cause people’s
eyes to glaze over. The potential social, political, and
economic effects of data integration are not exciting at
first glance, yet a closer look reveals the important impact
data integration has on how we construct geographical
reality in information systems. Every time information
about the world is recorded, classified, scaled, or
modelled, selections about what is important have been
made. The process of selection creates a representational
reality, expressed through data models and maps, that
influences decisions in domains from health care to
subsurface waste management. Policies that constrain or
enable a diversity of expression in databases ultimately
affect the construction of social and physical knowledge.
How data are classified and structured is a reflection of
political and social priorities, and these decisions have
implications for establishing policy priorities. Social
geographers, on the one hand, might agree that knowledge is socially constructed but, with few exceptions,
cannot pinpoint these influences in a technical context
(Demeritt 1996, 2001; Hacking 1999). By contrast,
such influences may be implicitly acknowledged in the
technical realm of GIS interoperability research, but
explicit reference to them is absent from methodologies
and technical literature.
There is a gap between scholars who focus on social
implications of data structuring (of which interoperability
is one form) and those who develop technical implementations for data interoperability. Social scholars of
science have done an excellent job of historicizing
classification, standardization attempts, and interoperability (Bowker 2000; Hacking 1999; Waterton 2002).
They have drawn attention to the vagaries of each of these
processes and the extent to which they reflect current
cultural context. Terminology, categories, and the bases
for classification are socially negotiated (Bryant 2000;
Hacking 1999; Mitchell 2001). The certainty that
classification, standardization, and interoperability are
imperfect methods of structuring information should not
paralyse us, nor should it constrain the development of
much-needed criteria and methods for merging data from
multiple sources. Indeed, contemporary research on
interoperability within GIScience seeks to find ways
of integrating spatial attributes based on their ‘‘true’’
meaning and context – though specifications do
frequently allow provision for extending standards to
comply with user needs (OpenGIS Consortium 1999).
This article will tease out intersections between social and
technical approaches to interoperability by addressing
social or theoretical issues that have clear and discernible
implications for semantic interoperability in GIS. Three
areas in particular will be discussed: (1) classification
48
and scale; (2) data models; and (3) governmental data
and interoperability policies. The article begins, necessarily, with an explanation of interoperability as a research
area in GIScience.
What is interoperability, and why does
it matter to geographers?
Interoperability is the pursuit of a means of combining
data sets based on a common spatial grammar. It is a
rubric for negotiation between systems and information
at many levels of GIS, including systems, syntax, and
semantics. Semantics are shorthand for attributes
connected with geographical features or objects, but
they also incorporate a worldview, an interpretation of
reality, that is variable and difficult to pinpoint even
within a single domain. Semantics also have to do the
extra work of describing multiple phenomena that share
the same classification or category. The term ‘‘interoperability’’ refers to the ability to share data and analysis
between multiple users on different platforms. Yaser Bishr
(1998) identifies six components of interoperability:
network protocols, hardware and OS, spatial data files,
database management systems, data models, and application (domain) semantics. Michael Goodchild, Max
Egenhofer, and Robin Fegeas (1997) have identified
eight pertinent elements. The dimensionality of interoperability, however, is not the chief focus here. Rather,
our focus is the belief that conformity among these
components will enable seamless sharing of data across
jurisdictions and borders. Though this is changing, a sense
that the interoperability problem can be ‘‘solved’’ through
conscientious programming efforts has prevailed
(Cuthbert 1999; Laurini 1998; Sheth 1999; Vckovski
1999b). More recent work emphasizes the dynamism of
language and the need for interoperability research to
develop methods for self-updating of semantic meaning –
through automated solutions (Kuhn 2001, 2003;
Mennis 2002).
Interoperability promises greater conceptual as well as
software stability for GIS (Bishr 1998; Brodeur and others
2003). Perceived benefits include reduced costs and time
to transform and manage data. The ‘‘existence of clear
principles for the internal logical consistency of GIS
databases’’ is perceived as a benefit of interoperability
(Goodchild and others 1997). Geodata standards are
touted as an ‘‘economic weapon’’ that will lead to more
open global markets (Salgé 1999). In this scenario,
standardized inventory systems will lead to frictionless
trade between nations. GIS interoperability promotes
shared organizational and analytical structures and
promises to make application development easier and to
increase the lifespan of applications. Like the open
architecture plan, introduced by IBM with the first PCs
cartographica (volume 40, issue 4)
Robin Fegeas
(1997) is not
listed. please
check
[Ver: A3B2 8.07r/W] File: {UTPress}Cartographica/CARTO-40(4)-005.3d
Creator: / Date/Time: 8.11.2005/10:38am Page: 47/62
Social Perspectives on Semantic Interoperability: Constraints on Geographical Knowledge from a Data Perspective
in the 1980s, and the current promise of Linux,
interoperability harbours the aspiration that common
specifications will seed and enhance the market. In
principle, interoperability will allow data to be used
across platforms without any loss of integrity (Kemp and
Vckovski 1998). It also promises to reduce the costs of
data acquisition and management. According to proponents, interoperability is a means of encouraging communication and information sharing across communities of
discourse (Schuurman 2002).
Interoperability between systems and data is a goal that
has been recognized since the 1970s (Arctur and others
1998) but pursued aggressively only since the early 1990s.
The 1994 US presidential order that brought the Federal
Geographic Data Committee (FGDC) into being was the
harbinger of a new era of analysis (Guptill 1994). The
subsequent proposal for the Spatial Data Transfer
Standard (SDTS) was intended as a means of implementing spatial data sharing. Other countries followed suit.
In Canada, for instance, Geoconnections was established
in 1999 by the federal government, along with industry
partners, to oversee the development of the Canadian
Geospatial Data Infrastructure (CGDI). As one of its five
policy thrusts the CGDI is to oversee the development of
geospatial standards. Its policy goals are closely aligned
with those of international spatial data standards, so that
Canadians can share information with other nations and
Canadian businesses can sell geospatial information
technology and services in the global marketplace.
Problems associated with standardizing data from disparate sources – each with its own set of institutional
priorities – have been acknowledged at a policy level,
yet a general belief persists that interoperability holds
the key to a future of borderless data (Albrecht 1999;
O’Donnell and Penton 1997; Tosta 1997).
Optimism has generally marked early forays into
standardization and interoperability for GIS; integration
standards are hailed as a means of allowing information
products to compete internationally. Like so many
scientific and technical problems, the logistics of the
enterprise have proved more difficult than was initially
imagined (Vckovski 1999a). Early researchers assumed
that the chief problems of interoperability would revolve
around data structuring techniques and networking
protocols. Ironically, it is semantic interoperability that
has proven the most difficult (Vckovski 1999a). Semantics
are the language in which variables and attributes are
expressed in databases, and they are subject to all the
vagaries of communication that have plagued linguists
and social theorists for decades (Haraway 1994; Kitcher
1998; Kwan 2002; Poster 1996; Sardar 2000). The response
among GIS and systems researchers has been to focus on
technical aspects of the problem. Interoperability has been
treated like a systems engineering problem.
cartographica (volume 40, issue 4)
Traditional Research Paradigms and the Algorithmic
Mantra
The terms ‘‘interoperability’’ and ‘‘standardization’’
have sometimes been used interchangeably, though they
should be differentiated. There are multiple visions of
interoperability (Bishr 1998; Sheth 1999). There is general
agreement, however, that interoperability is a broad
rubric and semantic standardization is one component.
Semantic standardization has also been recognized as one
of the most difficult aspects of interoperability to resolve
(Bishr and others 1999; Kottman 1999; Vckovski 1999a).
This article focuses primarily on semantic standardization
while accepting that different data schematics associated
with classification and data models and structures affect
semantics (Bishr 1998).
Semantic standardization is, broadly, the problem of
calling the same thing different names or slightly different
things the same name. Most terms are contextual:
‘‘urban’’ means different things depending on the place
and time the term is used. Different disciplines give
similar entities different names or give the same names
to different entities. The word ‘‘range,’’ for instance, can
refer to scope; the habitat of animal; a stove top used for
cooking; a spread of values; a series of mountains in a line;
a place where shooting is practised; or a verb meaning
‘‘to roam.’’ Even a semantic term as banal as ‘‘road’’ can
be understood very differently in different institutional
contexts. The Ministry of Forests and the Ministry of
Sustainable Resource Management in British Columbia
use the same provincial ‘‘feature code’’ dictionary to
define what constitutes a ‘‘road.’’ Yet the interpretation of
this feature is dramatically different, to the extent that
1:20,000 maps of the same area illustrate non-congruent
road networks (Schuurman 2002). Barry Smith and
David Mark (1998) pose the quintessential semantic
interoperability question: Where does the mouth of the
volcano end and the mountain start? In the discipline of
geography, ideas such as ‘‘cultural space’’ and ‘‘knowledge
maps’’ employ a complex and non-linear transformation
to the physical space, and the two are ultimately
‘‘unmappable’’ (C.J. Keylock, personal communication,
2 March 1999). This lack of semantic congruity leads
to misunderstandings even between closely related groups
of researchers and data users.
Science does attempt to mitigate this problem by insisting
on explicit definitions and standardization. The problem
with this approach is that nailing down meaning is
fraught. Negotiation is involved in the development of
categories. The medical category of ‘‘low birth weight,’’
for instance, is differently defined in multiple jurisdictions
precisely because it is a contentious category with
implications for statistics on the efficacy of national
health care systems. The politics of classification, however,
49
[Ver: A3B2 8.07r/W] File: {UTPress}Cartographica/CARTO-40(4)-005.3d
Creator: / Date/Time: 8.11.2005/10:38am Page: 47/62
Nadine Schuurman
are forgotten once the system is in place (Bowker 2000).
The process of formalization used in information
technologies attempts to tie users to precise categories.
From a user perspective, this can entail determining
whether the terms ‘‘clay’’ and ‘‘at risk’’ mean the same
thing in a New Brunswick database as they do in a
British Columbia (BC) database. The standardization
of geospatial data is, in theory, a means of ensuring
that databases are versatile and transferable between
applications and institutions. It does, however, restrict
the meaning of certain terms (e.g., ‘‘range,’’ ‘‘road,’’ or
‘‘low birth weight’’) in the interests of ensuring a common
‘‘language’’ for geospatial databases and thus expanding
the scope of geographic enquiry. Semantic diversity,
or the range of discourse that can be represented in a GIS,
is linked to the categories that are supported and their
complexity. Thus the way that semantic components of
databases are structured or ordered has a direct impact
on GIS. How a certain characteristic is defined (and
interpreted) has real effects on the resulting spatial entities
used for queries and display. Users of data tend to
interpret semantics based on their own history and
context, rather than on the context in which the data
were collected.
Scientists search for accuracy in the classification of every
phenomenon of study. Qualitative social scientists, likewise, conduct ethnographic and idiographic studies in
order to avoid voicing over differences muted by
totalizing theories or narratives. Geography has hosted a
number of debates about the suitability of nomothetic
versus idiographic research (Hess 1997; Walton 1995),
and there are clear divisions in the discipline over which
approach is more appropriate. Standardization is stifling
for some researchers. It is perceived as using a legalistic
and mechanistic approach to interoperability and data
sharing. Each information community, such as geology or
population health, has a vocabulary. Each has rules and
classification systems to create categories that organize
that vocabulary. Success comes from developing standardization in a manner that is acceptable to members of
specific intellectual and user communities.
There has been considerable research in GIScience on
ways of achieving semantic interoperability, but most of
this research has focused on technical solutions (Bishr
1998; Kottman 1999; Laurini 1998). Several strands of
technical interoperability and standardization research
in GIS are relevant to the problem of semantic
standardization. Each is united by an algorithmic
approach to semantic standardization. One such
approach focuses on federated data sharing environments
as one means of facilitating semantic standardization
(Abel and others 1998; Devogele, Parent, and Spaccapietra
1998; Sheth 1999). In one iteration of this model,
a mediator shares data among federated databases,
supporting scalable information and multiple ontologies
50
(Sheth 1999). Another version of the mediated system is
proposed by Frederico Fonseca and others (2002), who
envisage a GIS architecture for enabling integration of
semantic terms that incorporates ontologies. In this
instance, ontology is understood as a component of a
database, with different ontologies constituting different
classes (Fonseca and others 2002). The advantage of this
ontology-based system is that it surpasses the spatial
or map metaphor that has dominated other technical
approaches to spatial data.
Another tactic finds semantic similarities between data
sets. There are several ways to achieve this, but each
approach is linked by the fundamental goal of identifying
entities with similar descriptions in different databases
(Kashyap and Sheth 1996). One strategy is to develop
rules that stipulate conditions for inclusion in a particular
semantic category (Stock and Pullar 1999). Another
linguistic approach uses the principle of ‘‘semantic
proximity’’ between spatial objects. A more pragmatic
method is being developed by the Open GIS Consortium,
an affiliation of universities, industry, and government
partners. Open GIS uses existing technologies that can be
harnessed to facilitate data sharing using a component
architecture approach (Cuthbert 1999; Kottman 1999;
Voisard and Schweppe 1998). The advantage of designing
interoperability around distributed computing platforms
is that components can be designed to work between
existing software environments (Hardie 1998; Ungerer
and Goodchild 2002). A major advantage of Open GIS is
that it has the support of multiple institutions, an
important social influence in the implementation of
interoperability. To date, however, Open GIS efforts have
not focused specifically on semantic interoperability.
Most of these technically oriented solutions to semantic
heterogeneity arise at the intersection between computing
science and GIS. Certainly this research has contributed to
increased awareness of the problems of integrating
semantically heterogeneous data, but it is based on the
assumption that the problems with semantic heterogeneity arise from technical constraints and remain unrelated
to classification, scale, or policy. There is general
recognition that communities of discourse exist, but it
has been difficult to develop means of representing them
while enabling automated data sharing (Bishr and others
1999; Bittner and Edwards 2001; Kuhn 2003; Sheth 1999).
Technical approaches would be considerably strengthened
by closer attention to institutional practices that contribute to the development of unique communities of
discourse – which are subsequently embedded in data
semantics. Until very recently, a disjuncture has existed
between semantic standardization theory and the on-theground reality of navigating local data collection,
reporting, and institutional and governmental practices.
Research within the last five years has increasingly
recognized unique and separate communities of
cartographica (volume 40, issue 4)
[Ver: A3B2 8.07r/W] File: {UTPress}Cartographica/CARTO-40(4)-005.3d
Creator: / Date/Time: 8.11.2005/10:38am Page: 47/62
Social Perspectives on Semantic Interoperability: Constraints on Geographical Knowledge from a Data Perspective
discourse, or distinct geo-ontologies (Fonseca and
others 2002; Harvey 2005; Raubal 2001). Recognition
of distinct cultures of language use and attempts to
find ways to integrate their semantics emphasize the
extent to which GIScience has recognized the intersection
of philosophical and computational issues, but there
remains a gap between these discussions and successful
implementation of interoperability solutions. The gap
demonstrates that we are at the beginning of an
innovation curve in interoperability research; this permits
space to insert further social and political considerations
into the mix.
This article is a preliminary attempt to identify and
elaborate upon select social and political aspects of
interoperability, especially with respect to semantic
standardization. As such, it is an attempt to broaden
discussions about interoperability while invigorating them
with concerns that may be ignored in a singularly
technical research milieu. It incorporates a unique
strategy, however, in recognizing that all social problems
associated with GIS are ultimately technical in their
solution: there is no division between the two domains
(Schuurman 2004). The social and scientific have
never been separate; there has only been a refusal to
acknowledge their relationship (Latour 1993).
Interoperability challenges the traditional division
between social and technical realms, partly because
locating social influences requires some understanding
of the technical.
Show Me the Social
Research to date on interoperability (which includes
standardization) has been concerned with the coordination of applications and systems (Bishr and others 1999;
Buehler and McKee 2000; Egenhofer 1999b; Gahegan
1999; Kottman 1999; Sheth 1999; Vckovski 1999a, 1999b).
This approach is attractive and practical for software
engineers, given the requirements for formalization
associated with digital systems. It cannot account,
however, for a host of social issues as well as political
players involved in the process of creating specifications
and gaining support for interoperability. At the same
time, interoperability research has failed to attract a wide
constituency across geography and other research areas; it
remains the domain of technical researchers as opposed to
social scientists. As a result, technical standardization
specifications and protocols at the semantic level may be
established in the absence of consensus from many GIS
users who will later be concerned about representational
effects. Speaking to both social and technical issues
requires a hybrid approach that straddles human
geography and GIS.
In the past decade, human geographers have critiqued
GIS across a range of issues including epistemology,
cartographica (volume 40, issue 4)
militarism, and ethics (Pickles 1993, 1997; Sheppard 1993,
1995; Smith 1992). Many of these critics were legitimately
concerned with the prospect of a GIS that ignored social
dimensions, but they were unable to articulate their
apprehensions in a manner that allowed GIS researchers
to address the issues. Geographers concerned with social
issues in GIS have only recently started to communicate
meaningfully with researchers in GIScience (Sheppard
2005). These efforts have coincided with a period of
increasing recognition for science and technology studies
(STS), which highlight interaction between science and
society. STS researchers ask for recognition of the
complicated interplay between social and scientific players
in fashioning scientific and technical realities. Recognition
of how and when GIS problems are social is crucial for
developing appropriate technical solutions (Schuurman
1999).
In order to develop technical solutions for interoperability, one must understand the politics that were
involved in its negotiation. Francis Harvey (1999a) uses
the concept of boundary objects to describe the relationship between politics, technology, and society in fashioning specific iterations of technology. Boundary objects are
those that are recognized by more than one group, though
they may have different semantic designations. They act
as communication emissaries between groups (Harvey
1999b). Such social apparatuses are key elements in
illustrating social dimensions of GIS. Recognition that a
GIS concept or artefact is created through social or
political influences should not detract from its importance. Indeed, it should structure or instruct technical
solutions.
The most difficult and broadest of these is the issue of
classification, as it relates both to scale and data models.
Domain is usually regarded as the chief influence in
the development of categories, but there are multiple
variables involved, some with particular implications
for geospatial data. Related to classification is scale.
Indeed, the two are inseparable: as scale shifts, so do
classification categories. Though a number of human
geographers have begun to examine the effects of scale in
analysis (Marston 2000; Sheppard and McMaster 2004;
Towers 2000), interoperability researchers have largely
failed to include scale as an obstacle to integrating data
between domains. Interoperability stabilizes not only
classification categories but also data models – or the
digital representation of categories. Semantic categories
are partly shaped by data models, but more importantly
they are the expression of the range of ontologies
permitted by specific data models. There are, for instance,
specific ontological repercussions associated with objectoriented data models (Burrough and Frank 1996;
Couclelis 1996). This is a fair litany of social implications
for standardization and interoperability, and this article
only begins to identify them.
51
[Ver: A3B2 8.07r/W] File: {UTPress}Cartographica/CARTO-40(4)-005.3d
Creator: / Date/Time: 8.11.2005/10:38am Page: 47/62
Nadine Schuurman
Semantic Heterogeneity and the Classification
of Spatial Data
The most difficult aspect of data integration is not based
on network protocols or database schematics (Bishr 1997;
Harvey 1999a). Rather, it is the difficulty of creating
equivalence between attributes or characteristics of spatial
data (Vckovski, Brassel, and Schek 1999). In the
Vancouver region, for example, air quality is reported
using a numerical index. The index is a composite of
types of atmospheric pollution, but it is formulated
differently for each station. Figure 1 illustrates the
difficulty of comparing index figures for different
points in the metropolitan region, as they are not based
on the same pollutants. The casual user may not
realize that such data are not equivalent, and little
guidance is provided. Databases seldom include detailed
meaning about either attributes or their intended
use. Even when attribute interpretation seems selfevident,
semantic
interpretation
varies
widely
(Schuurman 2004).
Semantic interpretation often assumes a commensurability between attributes that is not justified. As a result, data
are merged based on irreconcilable meanings. Conversely,
researchers may exclude data based on an assumption that
differently spelled or articulated fields are not equivalent.
Figure 2 illustrates a well-log data file. Note the multiple
terms for clay. Automated integration systems flounder
when faced with such data.
These problems are exacerbated by the curt, often
truncated field descriptions typical of institutional
databases where meaning is implicit. Cryptic field headers
such as ‘‘%F’’ are not uncommon. Disparities in
interpretation of semantics lead to error and uncertainty
in analysis that is not easily detected by non-GIS experts
or even by data analysts. Semantic incommensurability is
illustrated by the use of the word ‘‘pond’’ in
Newfoundland and Labrador to refer to a large lake.
In mainland Canada, ‘‘pond’’ refers to a small, often
undrained body of water while ‘‘lake’’ refers to larger
bodies of water (Mark 1993). More complex attributes
invite even greater misinterpretation. These issues extend
to all disciplines. Population health researchers who work
across jurisdictions, for example, must deal with the
problem of seemingly identical categories being variously
interpreted. Categories such as ‘‘at risk’’ and ‘‘low birth
weight,’’ for example, are defined differently in each
Canadian province, making interprovincial comparison
difficult.
There have been a number of entreaties among interoperability researchers to establish a standard language
that limits semantic ambiguity (Frank and Kuhn 1999;
Vckovski 1999a; Vckovski, Brassel, and Schek 1999).
Indeed, the desire for a more tightly defined vocabulary –
semantic structures compatible with computer
architecture – has been expressed in various corners of
GIS (Hornsby and Egenhofer 1998; Kuhn 2001; Ruas
1998a, 1998b; Weibel and Dutton 1998). The reality of
Air Quality Monitoring Station
Air Quality Monitoring Station
Monitoring Station Identity
Number: T001
Location: Downtown Vancouver, Robson and Hornby St.
Latitude: N 49°16'58''
Longitude: W 123°07'14''
Elevation: 56m
Date: March 20 2003
Monitoring Station Identity
Number: T020
Location: Pitt Meadows, 188477 Dewdny Trunk Rd.
Latitude: N 49°14'43''
Longitude: W 122°42'33''
Elevation: 20m
Date: March 20 2003
Air Quality Index
Air Quality Index
Time
AQI
Air Quality
Time
AQI
00:00
6
Good
00:00
17
Good
06:00
12
Good
06:00
19
Good
12:00
15
Good
12:00
17
Good
18:00
11
Good
18:00
18
Good
Time
OZ
CO
NO2
SO2
Time
OZ
CO
NO2
SO2
PM10
00:00
5
0
6
0
00:00
17
1
1
0
8
06:00
12
0
2
0
06:00
19
0
0
0
6
12:00
15
0
5
0
12:00
17
0
0
0
6
18:00
11
0
5
0
18:00
18
1
1
0
6
Air Quality Sub-Index
Air Quality
Air Quality Sub-Index
Figure 1. Air quality indices for two monitoring stations in the Greater Vancouver area. Note that the Pitt Meadows station
includes a fifth attribute (PM10) in its calculation, yet both stations publish figures linked to the same index. Variation in
attributes is common for all monitoring stations in the area.
52
cartographica (volume 40, issue 4)
[Ver: A3B2 8.07r/W] File: {UTPress}Cartographica/CARTO-40(4)-005.3d
Creator: / Date/Time: 8.11.2005/10:38am Page: 47/62
Social Perspectives on Semantic Interoperability: Constraints on Geographical Knowledge from a Data Perspective
domain influence on semantics, however, has tempered
this goal. A tension remains between proponents of
strict vocabularies and those who view this ambition
as folly. Researchers in favour of building a vocabulary
that is consistent with digital representation argue
that our semantic structures have been too loose to
allow efficient representation or generalization of
objects and their complex interrelationships. They
contend that in order to translate meaning and
methods into digital terms, we need a stricter vocabulary
that allows us to impart contextual knowledge.
This maxim stresses that it is necessary to move to a
semantic system that conforms to machine architecture.
Other researchers have been at pains to build semantic
bases for operations and geographic phenomena based
on the argument that the traditional representation of
digital data as (x, y), plus a semantic code, has limited
the amount of contextual knowledge available for
making generalization and other decisions about
the interrelationships between objects (Hornsby and
Egenhofer 1998; Ruas 1998b). Strict semantic systems
are fiercely contested because of their premise that
language can be absolute.
Dianne Richardson, a senior researcher at the Canada
Centre for Remote Sensing (CCRS), bases her opposition
to rigid classification on a Canadian experience with
semantic standardization:
The conclusion that has been drawn, at least here in Canada,
is that it is very difficult to impose some kind of standard
for semantic classification . . . . In some form or another, the
actual semantic classifications need to be translated and
you can do it through data dictionaries or look-up tables
for an alternative approach. Now, an example is in my own
work where I have come up against that problem particularly
in transportation networks, hydrography, land-use coverage.
Canada as a whole spent a huge amount of time and money
trying to standardize the classifications so that the semantics
associated with it were the same across all the provinces.
That has just wasted many people’s time, a lot of money,
and a lot of effort . . . . So rather than investing a lot of time
on getting agreement on semantics. Good heavens, it is
just such an extraordinary task. Why would people waste
their time on it? (D. Richardson, personal interview,
26 October 1998)
Richardson’s view is gaining support, especially as
evidence accumulates that rigorous semantic structures
fail to enhance communication and data sharing (Brodeur
and others 2003; Harvey 2003). An increasing number of
researchers in GIScience have focused on incorporating
Figure 2. Well-log data table illustrating the variation in descriptions of comparable, contiguous materials.
cartographica (volume 40, issue 4)
53
Harvey 2003 is
not listed in the
reference list.
please check
[Ver: A3B2 8.07r/W] File: {UTPress}Cartographica/CARTO-40(4)-005.3d
Creator: / Date/Time: 8.11.2005/10:38am Page: 47/62
Nadine Schuurman
multiple ontologies with their own semantic usage into
GIS (Bittner and Edwards 2001; Brodaric and Hastings
2002; Cruz and others 2002; Fonseca and others 2002;
Frank 2001; Kuhn 2002; Raubal 2001; Smith and
Mark 1998; Winter 2001). Acknowledgement is emerging
among some interoperability researchers that there will
never be a standard set of semantics deployed for database
entry (Brodaric and Hastings 2002; Forer 1999). Instead,
researchers will have to work with a federated database
system in which translation between databases from
different domains is mediated at a meta-level (Abel and
others 1998; Sheth 1999). Virtual or temporary databases
will be generated through the aid of an administrator.
Though there are high hopes for a semi-automated system
(Devogele and others 1998), current practice involves
building translation tables in a painstaking way that
creates equivalences between every semantic designation
in every database. Translators in a federated database
system will continue to face the same problem that
confronts those who try to create a unified vocabulary.
Moreover, automated translators are notoriously cumbersome and narrow in scope.
Classification can be described as ‘‘a spatial, temporal or
spatio-temporal segmentation of the world . . . A classification system is a set of boxes, metaphorical or not, into
which things can be put in order to then do some kind of
work – bureaucratic or knowledge production’’ (Bowker
and Star 1996). Classification systems are, by definition,
not complete. They aim to include a ‘‘statistically
significant’’ level of occurrences of a given phenomenon.
Depending on the resolution of the study, different
phenomena will constitute a statistically significant
category. Classification remains a domain-specific undertaking, and one fraught with social and political influence.
The impetus for important standards is the desire for
global analysis of spatial data. Thus, it is important to
recognize that decisions about classification are decisions
about the extent of study areas. Clearly, the higher
the resolution of the study, the greater the number of
categories that emerge. Both scientific and social pressures
bear on the choice of scale associated with research.
Scale has been a minor concern for researchers examining
interoperability, in part because emphasis has focused on
more immediate problems (Arctur and others 1998). The
problem of semantic heterogeneity is essentially a problem
of scale, not just, as is generally assumed, one of domain.
The damming of the Yangtze River in China provides a
metaphor for the relationship between scale, classification,
and standardization. It also clarifies the scale-dependent
nature of naming and domain knowledge. The dam will
obliterate many small villages, agricultural areas, subsistence farms, and bicycle and footpaths to enable hydroelectric production and limit flooding. There is
understandable opposition from those whose livelihoods,
homes, and fields will be flooded, just as state bureaucrats
54
and engineers envision a great power source to solve
China’s energy problems. Similarly, the question in GIS
is how to retain those local knowledges and artefacts
while allowing the larger project to go ahead.
The translation between classification and standardization
is fundamentally an issue of scientific and governmental
communication (Bowker 2000). Not unlike nation-states,
scientific and social domains have their own conceptual
models. Certain domains may be better able to inscribe
their needs on emerging specifications from government
and industry. Geology, for instance, uses threedimensional surfaces. Each community of interest will
need to appeal to have its domain models incorporated
into the new vocabulary, and three-dimensional data,
for instance, will not be everyone’s priority (Jacquez
1999). Geologists will petition to have the standard
language include geologic terms and processes. But the
ontologies of all communities need to be addressed
(Peuquet, Smith, and Brogaard 1998). Clearly this is an
example of the merging of the political with the technical.
And, as in every political negotiation, there will be
winners and losers. Interoperability does not serve
everyone equally; it serves some knowledge domains
better than others. For instance, models and terms used in
natural-gas exploration are standardized by the petroleum
industry using a well-established and accepted data model
(PPDM); but community mapping groups (re)invent a
vocabulary with every project.
The creation of classification systems for data is, in effect,
the creation of a standardization infrastructure, a process
mediated by political pressures as well as scientific
motivations. Little research in geography has been
conducted on how raw data get converted into database
fields, but politics often determine what will be recorded
in the database (and what is excluded). As a result,
systems may become data driven rather than process
driven. Many women physicians, for example, spend
more units of time with each patient than their male
counterparts. As a result, they see their patients less
frequently but for longer intervals (Bowker and Star
1996). A data-driven classification system cannot account
for relationships outside of the classification system; it
may ostensibly appear that women are less efficient
physicians than men. An example from my own research
with urban geological datasets indicates that subsurface
groundwater records contain information about lithology
but not about the relationships between subsurface layers
or the processes that affect (shape) multiple layers. Field
notes associated with these processes are frequently
written in the record margins, but they do not become
part of the official record: that information is lost in the
analysis. A process-centred model would be better able to
model relationships among data. A key goal for semantic
standardization is the retention of information related to
processes associated with data. Even if individual domains
cartographica (volume 40, issue 4)
[Ver: A3B2 8.07r/W] File: {UTPress}Cartographica/CARTO-40(4)-005.3d
Creator: / Date/Time: 8.11.2005/10:38am Page: 47/62
Social Perspectives on Semantic Interoperability: Constraints on Geographical Knowledge from a Data Perspective
are able to record and incorporate information about
process, the utility of that information is limited by the
choice of data models.
Ontologies and Data Models: An Argument
against Inflexible Semantic Standardization
Ontology researchers over the past decade have recognized that the data models we are forced to choose
between represent very different ways of seeing the world
(Campari 1991; Couclelis 1992, 1996; Frank 1996; Kemp
and Vckovski 1998; Mark 1993; Peuquet and others 1998;
Rodriguez, Arnada, and Navarro 1996; Schatzki 1991;
Smith and Mark 1998; Volta and Egenhofer 1993). Object
data models epitomize the idea of discrete, separate
entities in a frictionless neutral space. In the process of
their definition, objects make a strong ontological
commitment. Field models, on the other hand, discretize
areas of geographical space but make a lesser ontological
commitment (Couclelis 1999). Both fields and objects
can be represented using either raster or vector, though
objects tend to be associated with vector representation
and fields with raster. Both field and object models
ultimately rely on a Newtonian view of the world coupled
with Euclidean geometry; neutral space has traditionally
been assumed (Couclelis 1999). This view of space,
however, is increasingly being questioned (Massey 1999;
Raper 2000). Neither fields nor objects allow the
characterization of complex interrelated geographic
entities or processes (Couclelis 1996; Goodchild 1992).
Each is a simplistic characterization of a complex
geographical reality.
Ontologies, ideally, should not be restricted to the digital
primitives available in current GIS, but discussions of
ontologies have led inexorably back to existing data
models. ‘‘They substitute,’’ Egenhofer has commented,
‘‘for the fundamental elements of geographic thought’’ –
in a database model of the world (Egenhofer 1999a).
Discussion about ontologies is taking place precisely
because there is increasingly suspicion that the data
models that underlie our primitives are not ideal
surrogates for spatial objects and their relations.
Nevertheless, a recursivity between available data models
and the framework for exploring geographic ontologies
persists. I have argued elsewhere that the field/object
dichotomy in GIS speaks to a tradition of binarism rather
than to any inherent duality in the geographical world
(Schuurman 1999).
Indeed, fields and objects are self-reinforcing. Current
thinking is often caught up in a false binary of either/or.
These models are our current alternatives, and, as a result,
all conversations end up revolving around the two
conceptualizations of space as either a set of objects or
an infinite field. This conundrum has led to an extreme
view in which there is a duality to geographical space,
cartographica (volume 40, issue 4)
akin to the light and wave models of light used in physics
(Peuquet 1988). It would be more productive to envisage
current data models as provisional – in the absence
of more complete technologies for representing the
complexities of geographical features and relationships.
Ontological implications of data models have been widely
discussed in various sectors of the GIS community over
the past decade, but these conversations were absent until
recently from the interoperability literature. An exception
is Amit Sheth (1999), who argues that the present (third)
generation of interoperability research must support
multiple ontologies associated with different domains.
More recently, ontologies have been closely linked to
standardization. Smith and Mark (2001) illustrate that
non-experts conceptualize geospatial phenomena using
a simple, fairly consistent set of spatial categories.
The implication is that ontological terms such as
‘‘object’’ or ‘‘entity’’ hold common meanings that can
be naturalized. This naturalization of perception and
classification is problematic, however, because it is based
on a clear separation between ontology and epistemology
that is difficult to support, given that data models
are used in different domains, many with varying
epistemologies. One way to acknowledge that geospatial
data include various levels of inconsistency, as well as
varying reliability in terms of correspondence to the ‘‘real
world,’’ is to include a meta-layer of information about
the observer and observations that documents information about measurement techniques as well as subjective
influences on the data (Frank 2001). This tactic has been
attempted in the geosciences with the development of the
North American Geological Data Model (NADM).
NADM provides a means of documenting recognized
institutional epistemologies that are invariably manifest
in the choice of terms used to describe geographical
phenomena (Brodaric and Hastings 2002). NADM is
innovative in that it attempts to resolves the traditional
problems of separating ontologies and epistemologies.
The two are inevitably intertwined, and one of the major
contributions of post-structuralism has been to demonstrate this fact (Haraway 2000).
Even if terminology and processes are successfully
standardized, the risk remains that standards will be
altered after their imposition to suit various environments. Since the advent of the first shared standards
between US agencies, for example, intra-agency additions
and revisions have continually appeared as institutions
attempt to incorporate local context (Guptill 1994).
Data are information with contextual frameworks that
interoperability standards may fail to accommodate.
Institutions are quite skilled in altering standards to
suit their data and analysis needs – thus the dictum that
‘‘God must have loved standards; she made so many.’’
Jochen Albrecht echoes this irony in his summary article
on the multiple standards presently in use or under
55
Smith and Mark
(2001) is not
listed in the
reference list.
please check.
[Ver: A3B2 8.07r/W] File: {UTPress}Cartographica/CARTO-40(4)-005.3d
Creator: / Date/Time: 8.11.2005/10:38am Page: 47/62
Nadine Schuurman
construction (1999). Interoperability should support
different versions of the same situation, an approach
called ‘‘branching’’ (Peuquet and others 1998). The most
sophisticated model at present is the Spatial Archive and
Interchange Format (SAIF) (Albrecht 1999). Despite the
sophistication of this proposed standard (it supports 43
iterations of field and object data models), each iteration
of an event remains restricted by semantics and data
definitions.
The challenge for users and domains is to use the available
provisional data models in GIS to support standardized
categories without making immutable ontological
statements that are difficult to revise. Every ontology,
no matter how variegated and wide a set of categories
it uses, has limitations (Mulhall 1996).
Vanishing or burgeoning? Policy, Government,
and Interoperability
Depending on who is talking, national mapping organizations (NMOs), such as the United States Geological
Survey (USGS) or the Geological Survey of Canada
(GSC), are integral to achieving semantic standardization
or institutions that are losing their grip on data
production as privatization proliferates (Forer 1999).
In a trenchant analysis of changes in the role of NMOs,
Forer identifies that, from the perspective of the NMOs,
their role has never been more critical as they prepare to
lead their respective countries into a world of seamless
data interoperability. Depending on the country, of
course, the strategy is different. Canada, New Zealand,
and Australia sing the praises of cost recovery, while the
better proven strategy of the United States in providing
free data as a means of seeding the market is offered as the
democratic alternative (Mooney and Grant 1997;
Morrison 1997; O’Donnell and Penton 1997; Robertson
and Gartner 1997). Representatives of NMOs believe that
interoperability will emerge from cartographic and
bureaucratic traditions based on limited ontologies for
‘‘agreed upon geographic objects’’ at well-defined scales
(Forer 1999, 3). This view fails to take into account
the changes wrought by digital data, which are fluid,
unstable, and mobile – indeed, they share many of
the qualities of global capital. Forer argues that NMOs
will, even in an optimistic scenario, retain only the
ability to influence the infrastructure upon which
data will accumulate. Data production itself will go the
way of globalization, becoming disparate and privatized.
In this scenario, unless NMOs rapidly establish the
basis for data infrastructures, they will be overtaken by
data producers of many stripes who may care little for
protocol and even less for the historical role of NMOs
in setting the standards for scale, ontologies, and formats
of data.
56
Case in Point: Bore-Hole Standards
The intersecting roles of the private sector, local
government, and NMOs are illustrated by the example
of bore-hole data, which are used for groundwater
modelling and environmental decision making affecting
the earth’s subsurface. In Canada and, to a lesser extent,
in the United States, bore-hole data are used to understand the lithological structure of the subsurface, and
especially to identify aquifers and aquitards (layers that do
not transport water). The location and extent of aquifers
and aquitards are important because they are used to
determine where it is safe to allow agricultural operations
and where to bury waste or store it in landfills, as well as
in many other environmental decisions. Most of the data
used for these important decisions come from driller
reports from private firms.
In some Canadian provinces, it is mandatory for drillers
to submit reports after drilling any well. In others, driller
logs are submitted on a voluntary basis. Few drillers are
trained to identify geological materials, with the result
that the quality of well-log data varies widely. Moreover,
drillers may recognize rocks and other material but call
them different names. In a minority of cases, the province
has developed a mandatory classification system, but this
is rare. The example of British Columbia, where more
than 120 terms for lithological materials are found in the
well logs, is indicative of the extent of data heterogeneity.
Semantic standardization is necessary just to be able to
use data, and it is accomplished by creating fixed common
fields and designating data to those fields. The process
of identifying materials that will be grouped into a single
category is one of ontology building . This ontology
building, however, is not being undertaken by
Geoconnections, the Canadian federal body responsible
for spatial data infrastructure. Instead, isolated pods
of academics and government scientists are working with
the existing data and private drillers in an attempt to
rationalize lithological classifications locally (Logan,
Russell, and Sharpe 2001; Desbarats and others 2001).
There are nascent attempts to link entities of this
piecemeal network. The task itself is daunting, partly
because the motivations of interested parties vary widely.
Drillers submit their logs only as a courtesy, or because
they are mandated to do so. Provincial governments,
which control water resources in Canada, are uniformly
under financial pressure and are frequently reluctant to
commit resources to what might be viewed as an esoteric
task. Geoscientists are interested in the process of
lithological standardization because data are necessary
for models, but data quality is not always a pressing
concern. GIS specialists are concerned with data quality
and interoperability but less with any particular domain.
My own research has recognized the plurality of
approaches to lithological standardization, as well as its
cartographica (volume 40, issue 4)
[Ver: A3B2 8.07r/W] File: {UTPress}Cartographica/CARTO-40(4)-005.3d
Creator: / Date/Time: 8.11.2005/10:38am Page: 47/62
Social Perspectives on Semantic Interoperability: Constraints on Geographical Knowledge from a Data Perspective
ontological implications for the geosciences as well as
for GIS (Schuurman 2002). Coordinating multiple
stakeholders is one prong of a broader project to find
ways of integrating large, multiple-source data sets.
What is noteworthy, from a social and policy perspective,
is that it is likely that a national classification system for
subsurface lithologies will be developed and implemented
outside of a national spatial data infrastructure. The
repercussions of an organic classification that results from
an anarchic set of interested parties are varied. Lack of a
supporting national data infrastructure is one, though this
may be offset by the flexibility of possible approaches to
classification and standardization. It does point, however,
to the importance of social networks in developing
science.
What might an alternative look like? Database
Ethnographies and Ontology-Based Metadata
I have outlined the myriad challenges and obstacles to
semantic interoperability, with an emphasis on the
homogenizing effects of standardization, the inability of
policy-based national infrastructures to build viable
frameworks for interoperability, and the vagaries of
classification that lead to skeletal understandings of
complex phenomena. A very gloomy portrait. There are
alternatives, however. One is being developed by my
research group at the Institute for Health Research and
Education (IHRE) at Simon Fraser University. We are
developing an ontology-based metadata protocol for
population health data that builds upon ISO 19115.
Rather than include only traditional metadata, such as
lineage, scale, projection, boundary coordinates, meaning
of attribute codes, and data authorship, we are collecting
‘‘ontology-based’’ metadata that include measurement
techniques, rationale for collection, and usage within
specific communities of discourse. These data are being
collected from data stewards of existing population health
databases as well as from data contributors, using a
method of database ethnography.
Database ethnographies draw on research from science
and technology studies in which scientists and their
practices are interrogated in order to reveal the contingent
and contextual nature of scientific practices (Latour 1987;
Law 1994; Law and Hassard 1999). The initial pilot
project is being conducted with data from the Vital
Statistics branch of the Ministry of Health for the British
Columbia provincial government. Vital Statistics collects
mortality and morbidity data for BC based on the
International Classification of Diseases (ICD) 10 codes.
Our hypothesis, however, is that the ICD codes are used
very differently in the field than is understood by the
Ministry of Health. The codes are, in effect, boundary
objects that have different meanings in specific communities of discourse but act as semantic emissaries between
cartographica (volume 40, issue 4)
groups. Interviews with coroners and AIDS physicians as
well as street nurses are designed to corroborate this
hypothesis by illustrating that their use of the ICD
classification terms to categorize HIV/AIDS deaths varies
depending on their relationship with the patient, the
deceased’s family context, and the degree to which the
person responsible for assigning the category is comfortable with it. Varied usage of the same ICD categories
illustrates that semantics have unique ontological or use
contexts that form an important dimension of metadata.
Moreover, context-based metadata provide the basis for
making defensible decisions about data integration.
Clearly, it is not practical to reverse engineer all data to
include more expansive metadata. Recognition of the
value of dimensionalized metadata, however, is the basis
for including them in future data collection projects.
The database ethnographies project recognizes that
communities of use (e.g., population health researchers
in Canada) must be consulted and engaged with such
efforts in order to ensure success – and, indeed, to elicit
the necessary contextual information. Ontology-based
metadata are the most potent available venue for
contextualizing data and providing a basis for making
better decisions about data usage and interoperability.
Boundary objects are more clearly understood between
communities when they are contextualized.
Future Research: Semantic Heterogeneity in the World
of Interoperability
I have illustrated here that there is a range of concerns
about semantic standardization: (1) that semantic heterogeneity is as much a function of classification and scale as
of domain; (2) that there is a need to retain ontological
flexibility associated with semantic standardization; and
(3) that the infrastructural role of NMOs is shifting as the
privatization of data proliferates, and this will affect the
ontologies of geographical objects. Although there have
been a number of proposed and preliminary technical
solutions to semantic standardization, including the
stabilization of categories and federated databases, these
(1) do not necessarily include process information and
(2) often rely on exact classification, a set of well-defined
rule for membership, or a specific data model. This
critical approach to existing interoperability efforts should
be interpreted not as a voice against algorithmic solutions
to interoperability but as a plea for incorporation of social
and qualitative input – in the form of metadata. We need
to continue to solve technical problems while finding
ways to work with complex, local, and variable geographical data. Despite the cumulative caution expressed
thus far, interoperability and standardization remain the
best hope we have for encouraging analysis and scientific
cooperation at a global scale based on a growing pool of
geospatial data.
57
[Ver: A3B2 8.07r/W] File: {UTPress}Cartographica/CARTO-40(4)-005.3d
Creator: / Date/Time: 8.11.2005/10:38am Page: 47/62
Nadine Schuurman
Acknowledgements
I am indebted to Francis Harvey, who offered a number of
lucid insights into the complex relationships between
semantics, interoperability, and data infrastructures.
This research was partially funded by Social Sciences
and Humanities Research Council (SSHRC) grant #4012001-1497 and by a grant from the Canadian Institutes of
Health Research (CIHR), #116338. A much earlier draft
of this article was presented at a workshop on GIS in a
Changing Society coordinated by Duane Marble and
held at the Center for Mapping, Ohio State University,
in May 2001.
Author Information
Nadine Schuurman, Department of Geography,
Simon Fraser University, 8888 University Drive,
Burnaby, BC V5A 1S6 Canada. Tel.: (604) 291-3320.
E-mail: nadine@sfu.ca.
References
Abel, D.J., B.C. Ooi, K.-L. Tan, and S.H. Tan. 1998. ‘‘Towards
Integrated
Geographical
Information
Processing.’’
International Journal of Geographical Information Science
12: 353–71.
Albrecht, J. 1999. ‘‘Geospatial Information Standards:
A Comparative Study of Approaches in the Standardisation
of Geospatial Information.’’ Computer and Geosciences 25:
9–24.
Arctur, D., D. Hair, G. Timson, E. Martin, and R. Fegeas. 1998.
‘‘Issues and Prospects for the Next Generation of the Spatial
Data Transfer Standard (SDTS).’’ International Journal of
Geographical Information Science 12: 403–25.
Bishr, Y.A. 1997. ‘‘Semantic Aspects of Interoperable GIS.’’
PhD diss., International Institute for Aerospace Survey and
Earth Sciences (ITC), Enschede, The Netherlands.
Bryant, R. 2000. Discovery and Decision: Exploring the
Metaphysics and Epistemology of Scientific Classification.
London: Associated University Press.
Buehler, K., and L. McKee. 2000. ‘‘Effective Compilation and
Browsing of Geographic Metadata Using Common
Interoperability Tools.’’ In Conference of European
Statisticians.
Neuchatel,
Switzerland:
Statistical
Commission and Economic Commission for Europe.
Available at http://www.unece.org/stats/documents/2000/
04/gis/20.e.pdf.
Burrough, P.A., and A.U. Frank, eds. 1996. Geographic Objects
with Indeterminate Boundaries. London: Taylor & Francis.
Campari, I. 1991. ‘‘Some Notes on Geographic
Information
Systems:
the
Relationship
between
Their Practical Application and Their Theoretical Evolution.’’
In Cognitive and Linguistic Aspects of Geographic Space,
ed. D.M. Mark and A.U. Frank. London: Kluwer Academic.
35–44.
Couclelis, H. 1992. ‘‘People Manipulate Objects but Cultivate
Fields: Beyond the Raster–Vector Debate in GIS.’’ In Theories
and Methods of Spatial-Temporal Reasoning in Geographic
Space, ed. A.U. Frank, I. Campari, and U. Formentini. Berlin:
Springer-Verlag. 65–77.
———. 1996. ‘‘Towards an Operational Typology of
Geographic Entities with Ill-Defined Boundaries.’’
In Geographic Objects with Indeterminate Boundaries,
ed. P.A. Burrough and A.U. Frank. London: Taylor & Francis.
45–55.
———. 1999. ‘‘Space, Time, Geography.’’ In Geographical
Information Systems: Principles, Techniques, Management
and Applications, ed. P.A. Longley, M.F. Goodchild,
D.J. Maguire, and D.W. Rhind. New York: Wiley. 29–38.
Cruz, I.F., A. Rajendran, W. Sunna, and N. Wiegand. 2002.
‘‘Handling Semantic Heterogeneities Using Declarative
Agreements.’’ Paper read at ACM International Symposium
on Advances in GIS, 8–9 November, Mclean, VA.
———. 1998. ‘‘Overcoming the Semantic and Other Barriers to
GIS Interoperability.’’ International Journal of Geographical
Information Science 12: 299–314.
Cuthbert, A. 1999. ‘‘OpenGIS: Tales from a Small Market
Town.’’ In Interoperating Geographic Information Systems,
ed. A. Vckovski, K.E. Brassel, and H.J. Schek. Zurich: Springer.
17–28.
Bishr, Y.A., H. Pundt, W. Kuhn, and M. Radwan. 1999. ‘‘Probing
the Concept of Information Communities: A First Step toward
Semantic Interoperability.’’ In Interoperating Geographic
Information Systems, ed. M.F. Goodchild, M. Egenhofer, R.
Fegeas, and C. Kottman. Boston: Kluwer Academic. 55–69.
Demeritt, D. 1996. ‘‘Social Theory and the Reconstruction of
Science and Geography.’’ Transactions of the Institute of
British Geographers 21: 484–503.
Bittner, T., and G. Edwards. 2001. ‘‘Towards an Ontology for
Geomatics.’’ Geomatica 55: 475–90.
Bowker, G. 2000. ‘‘Mapping Biodiversity.’’ International
Journal of Geographical Information Science 14: 739–54.
Bowker, G.C., and S.L. Star. 1996. How Things (ActorNet)work: Classification, Magic and the Ubiquity of
Standards. Available at http://www.cse.ucsd.edu/users/
goguen/courses/268D/actnet.html.
Brodaric, B., and J. Hastings. 2002. ‘‘An Object Model for
Geologic Map Information.’’ Paper read at Joint International
Symposium on Geospatial Theory, Processing and
Applications, July, Ottawa, ON.
58
Brodeur, J., Y. Bedard, G. Edwards, and B. Moulin. 2003.
‘‘Revisiting the Concept of Geospatial Data Interoperability
within the Scope of the Human Communication Processes.’’
Transactions in GIS 7: 243–65.
———. 2001. ‘‘The Construction of Global Warming and the
Politics of Science.’’ Annals of the Association of American
Geographers 91: 307–37.
Desbarats, A.J., M.J. Hinton, C.E. Logan, and D.R. Sharpe.
2001. ‘‘Geostatistical Mapping of Leakance in a Regional
Aquitard, Oak Ridges Moraine Area, Ontario, Canada.
Hydrology Journal 9: 79–96.
Devogele, T., C. Parent, and S. Spaccapietra. 1998. ‘‘On Spatial
Database Integration.’’ International Journal of Geographical
Information Science 12: 335–52.
Egenhofer, M. 1999a. ‘‘The Emperor’s New Clothes.’’ Paper
read at the Annual Meeting of the Association of American
Geographers, 27 March, Honolulu, HI.
cartographica (volume 40, issue 4)
[Ver: A3B2 8.07r/W] File: {UTPress}Cartographica/CARTO-40(4)-005.3d
Creator: / Date/Time: 8.11.2005/10:38am Page: 47/62
Social Perspectives on Semantic Interoperability: Constraints on Geographical Knowledge from a Data Perspective
———. 1999b. ‘‘Introduction: Theory and Concepts.’’ In
Interoperating Geographic Information Systems, ed.
M. Goodchild, M. Egenhofer, R. Fegeas, and C. Kottman.
Boston: Kluwer Academic. 1–4.
Fonseca, F.T., M.J. Egenhofer, P. Agouris, and G. Camara. 2002.
‘‘Using Ontologies for Integrated Information Systems.’’
Transactions in GIS 6: 231–57.
Forer, P. 1999. ‘‘Guest Editorial: Whose Framework? Whose
World?’’ Transactions in GIS 3: 3–6.
Frank, A.U. 1996. ‘‘The Prevalence of Objects with Sharp
Boundaries in GIS.’’ In Geographic Objects with Indeterminate
Boundaries, ed. P.A. Burrough and A.U. Frank. Bristol, PA:
Taylor & Francis. 29–40.
———. 2001. ‘‘Tiers of Ontology and Consistency Constraints in
Geographical Information Systems.’’ International Journal of
Geographical Information Science 15: 667–78.
Frank, A.U., and W. Kuhn. 1999. ‘‘A Specification Language
for Interoperable GIS.’’ In Interoperating Geographic
Information Systems, ed. M.F. Goodchild, M. Egenhofer,
R. Fegeas, and C. Kottman. Boston: Kluwer Academic.
123–32.
———. 2005. ‘‘The Linguistic Trading Zones of Semantic
Interoperability.’’ In Re-presenting GIS, ed. P. Fisher and
D.J. Unwin. London: Wiley. XXX–XXX.
Hess, D. 1997. Science Studies: An Advanced Introduction.
New York: New York University Press.
Hornsby, K., and M. Egenhofer. 1998. ‘‘Identity-Based Change
Operations for Composite Objects.’’ Paper read at 8th
International Symposium on Spatial Data Handling, July,
Vancouver, BC.
Jacquez, G. 1999. ‘‘Ontology of Fields.’’ Paper read at the
Annual Meeting of the Association of American Geographers,
24 March, Honolulu, HI.
Kashyap, V., and A. Sheth. 1996. ‘‘Semantic and Schematic
Similarities between Database Objects: A Context-Based
Approach.’’ The VLDB Journal 5: 276–304.
Kemp, K., and A. Vckovski. 1998. ‘‘Towards an Ontology of
Fields.’’ Paper read at the Third International Conference on
GeoComputation, 17–19 September, Bristol, UK.
Kitcher, P. 1998. ‘‘A Plea for Science Studies.’’ In A House Built
On Sand: Exposing Postmodernist Myths about Science, ed.
N. Koertge. New York: Oxford University Press. 32–52.
Gahegan, M. 1999. ‘‘Characterizing the Semantic Content of
Geographic Data, Models, and Systems.’’ In Interoperating
Geographic Information Systems, ed. M.F Goodchild,
M. Egenhofer, R. Fegeas, and C. Kottman. Boston: Kluwer
Academic. 71–83.
Kottman, C. 1999. ‘‘The Open GIS Consortium and Progress
Toward Interoperability in GIS.’’ In Interoperating Geographic
Information Systems, ed. M.F. Goodchild, M. Egenhofer, R.
Fegeas, and C. Kottman. Boston: Kluwer Academic. 39–54.
Goodchild, M.F. 1992. ‘‘Geographical Information Science.’’
International Journal of Geographical Information Systems
6: 31–45.
Kuhn, W. 2001. ‘‘Ontologies in Support of Activities in
Geographical Space.’’ International Journal of Geographical
Information Science 15: 613–31.
Goodchild, M.F., M. Egenhofer, and R. Fegeas. 1997.
‘‘Interoperating GISs: Report of a Specialist Meeting Held
under the Auspices of the Varenius Project, December 5–6,
1997.’’ In Varenius Reports. Santa Barbara, CA: NCGIA.
Available at http://www.ncgia.ucsb.edu/conf/interop97/
report.html.
———. 2002. ‘‘Modeling the Semantics of Geographic
Categories through Conceptual Integration.’’ In Geographic
Information Science, ed. M.J. Egenhofer and D.M. Mark.
Berlin: Springer-Verlag. 108–18.
Guptill, S.C. 1994. ‘‘Synchronization of Discrete Geospatial
Data Bases.’’ In Advances in GIS Research: Proceedings of
the Sixth International Symposium on Spatial Data
Handling, vol. 2, ed. T.C. Waugh and R.G. Healey. London:
Taylor & Francis. 945–56.
Hacking, I. 1999. The Social Construction of What.
Cambridge, MA: Harvard University Press.
Haraway, D. 1994. ‘‘The Promises of Monsters: A
Regenerative Politics for Inappropriate/d Others.’’ In Cultural
Studies Reader, ed. L. Grossberg. New York: Routledge.
295–337.
Haraway, D.J. 2000. How Like a Leaf: An Interview with
Thyrza Nichols Goodeve/Donna J. Haraway. New York:
Routledge.
———. 2003. ‘‘Semantic Reference Systems.’’ International
Journal of Geographical Information Science 17: 405–9.
Kwan, M.-P. 2002. ‘‘Feminist Visualization: Re-envisioning
GIS as a Method in Feminist Geographic Research.’’ Annals of
the Association of American Geographers 92: 645–61.
Latour, B. 1987. Science in Action. Cambridge, MA: Harvard
University Press.
———. 1993. We Have Never Been Modern, trans. C. Porter.
Cambridge, MA: Harvard University Press.
Laurini, R. 1998. ‘‘Spatial Multi-database Topological
Continuity and Indexing: A Step towards Seamless GIS
Data Interoperability.’’ Geographical Information Science 12:
373–402.
Law, J. 1994. Organizing Modernity. Oxford: Blackwell.
Hardie, A. 1998. ‘‘Present State of Web-GIS.’’ Cartography
27/2: 11–26.
Law, J., and J. Hassard, eds. 1999. Actor-Network Theory and
After. London: Blackwell.
Harvey, F. 1999a. ‘‘Designing for Interoperability:
Overcoming Semantic Differences.’’ In Interoperating
Geographic Information Systems, ed. M.F Goodchild,
M. Egenhofer, R. Fegeas, and C. Kottman. Boston:
Kluwer Academic. 85–97.
Logan, C., H.A.J. Russell, and D.R. Sharpe. 2001. Regional
Three-Dimensional Stratigraphic Modelling of the Oak
Ridges Moraine Area, Southern Ontario. Geological Survey
of Canada Current Research Report No. 2001-D1. Ottawa:
National Resources Canada. Available at http://dsp-psd.
communication.gc.ca/Collection/M44-2001-D1E.pdf.
———. 1999b. ‘‘What Does Interoperability Mean?’’ Paper read
at the Annual Meeting of the Association of American
Geographers, 27 March 27, Honolulu, HI.
cartographica (volume 40, issue 4)
Mark, D.M. 1993. ‘‘Towards a Theoretical Framework
for Geographic Entity Types.’’ In Spatial Information
59
Please
update
page range.
[Ver: A3B2 8.07r/W] File: {UTPress}Cartographica/CARTO-40(4)-005.3d
Creator: / Date/Time: 8.11.2005/10:38am Page: 47/62
Nadine Schuurman
Theory: A Theoretical Basis for GIS, COSIT ’93, ed. A.U. Frank
and I. Campari. Berlin: Springer-Verlag. 270–83.
Experience.’’ In Framework for the World, ed. D. Rhind.
Cambridge, UK: GeoInformation International. 247–64.
Marston, S.A. 2000. ‘‘The Social Construction of Scale.’’
Progress in Human Geography 24: 219–42.
Rodriguez, F.H., G.B. Arnada, and A.M. Navarro. 1996. ‘‘An
Ontology-Based Approach to Spatial Information Modeling.’’
In Advances in GIS Research II: Proceedings of the Seventh
International Symposium on Spatial Data Handling, edited by
M.-J. Kraak, M. Molenaar, and E.M. Fendel. London: Taylor &
Francis. 801–13.
Massey, D. 1999. ‘‘Space-Time, ‘Science’ and the Relationship
between Physical Geography and Human Geography.’’
Transactions Institute of British Geographers 24: 261–76.
Mennis, J. 2002. ‘‘Derivation and Implementation of a
Semantic GIS Data Model Informed by Principles of
Cognition.’’ Computers, Environment and Urban Systems 27:
455–79.
Ruas, A. 1998a. ‘‘A Method for Building Displacement in
Automated Generalisation.’’ International Journal of
Geographical Information Systems 12: 789–803.
Mitchell, A. 2001. ‘‘Nature’s Family Tree Gets Axed?’’ The
Globe and Mail, 30 June 30: Focus.
———. 1998b. ‘‘OO-Constraint Modeling to Automate Urban
Generalization Process.’’ In Proceedings from the 8th
International Symposium on Spatial Data Handling, ed. T.K.
Poiker and N. Chrisman. Berlin: Springer. 225–35.
Mooney, D.J., and D.M. Grant. 1997. ‘‘The Australian National
Spatial Data Infrastructure.’’ In Framework for the World,
ed. D. Rhind. Cambridge, UK: GeoInformation International.
187–201.
Morrison, J. 1997. ‘‘Topographic Mapping in the Twenty-First
Century.’’ In Framework for the World, ed. D. Rhind.
Cambridge, UK: GeoInformation International. 14–27.
Mulhall, S. 1996. Heidegger and Being and Time. London:
Routledge.
O’Donnell, J.H., and C.R. Penton. 1997.‘‘Canadian Perspectives
on the Future of National Mapping Organizations.’’
In Framework for the World, ed. D. Rhind. Cambridge,
UK: GeoInformation International. 214–25.
Open GIS Consortium. 1999. ‘‘Topic 14: Semantics and
Information Communities.’’ In The OpenGIS Abstract
Specification, ed. C. Kottman. Wayland, MA: OpenGIS
Consortium. Available at http://portal.opengeospatial.org/
files/index.php?artifact_id ¼ 902.
Peuquet, D., B. Smith, and B. Brogaard. 1998.‘‘The Ontology of
Fields: Report of a Specialist Meeting held under the Auspices
of the Varenius Project.’’ Santa Barbara, CA: NCGIA. Available
at http://www.ncgia.ucsb.edu/Publications/Varenius_Reports/
Ontology_of_Fields.pdf.
Sardar, Z. 2000. Thomas Kuhn and the Science Wars.
Trumpington, UK: Icon.
Schatzki, T. 1991. ‘‘Spatial Ontology and Explanation.’’ Annals
of the Association of American Geographers 81: 650–70.
Schuurman, N. 1999. Critical GIS: Theorizing an Emerging
Science. Cartographica 36/4: Monograph 53.
———. 2002. ‘‘Flexible Standardization: Making Interoperability
Accessible to Agencies with Limited Resources.’’ Cartography
and Geographic Information Science 29: 343–53.
———. 2004. GIS: A Short Introduction. Oxford: Blackwell.
Sheppard, E. 1993. ‘‘Automated Geography: What Kind of
Geography for What Kind of Society?’’ Professional
Geographer 45: 457–60.
———. 1995. ‘‘Sleeping with the Enemy, or Keeping
the Conversation Going?’’ Environment and Planning A 27:
1026–28.
Peuquet, D.J. 1988. ‘‘Representations of Geographic Space:
Toward a Conceptual Synthesis.’’ Annals of the Association of
American Geographers 78: 375–94.
———. 2005. ‘‘Knowledge Production through Critical GIS:
Genealogy and Prospects.’’ Cartographica 40/4: XX–XX.
Pickles, J. 1993. ‘‘Discourse on Method and the History of
Discipline: Reflections on Jerome Dobson’s 1993 ‘Automated
Geography.’’’ Professional Geographer 45: 451–55.
Sheppard, E., and R.B. McMaster, eds. 2004. Scale and
Geographic Inquiry: Nature, Society, and Method. Oxford:
Blackwell.
———. 1997. ‘‘Tool or Science? GIS, Technoscience and the
Theoretical Turn.’’ Annals of the Association of American
Geographers 87: 363–72.
Sheth, A.P. 1999. ‘‘Changing Focus on Interoperability in
Information Systems: From System, Syntax, Structure to
Semantics.’’ In Interoperating Geographic Information
Systems, ed. M.F Goodchild, M. Egenhofer, R. Fegeas, and
C. Kottman. Boston: Kluwer Academic. 5–29.
Poster, M. 1996. ‘‘Databases as Discourse, or Electronic
Interpellations.’’ In Computers, Surveillance and Privacy, ed.
D. Lyon and E. Zureik. Minneapolis: University of Minnesota
Press. 175–92.
Raper, J. 2000. Multidimensional Geographic Information
Science. New York: Taylor & Francis.
Raubal, M. 2001. ‘‘Ontology and Epistemology for AgentBased Wayfinding Simulation.’’ International Journal of
Geographical Information Science 15: 653–65.
Robertson, W.A., and C. Gartner. 1997. ‘‘The Reform of
National Mapping Organizations: The New Zealand
60
Salgé, F. 1999. ‘‘National and International Data Standards.’’
In Geographical Information Systems: Principles, Techniques,
Management and Applications, ed. P.A. Longley,
M.F. Goodchild, D.J. Maguire, and D.W. Rhind. New York:
Wiley. 693–706.
Smith, B., and D.M. Mark. 1998. ‘‘Ontology and Geographic
Kinds.’’ In Proceedings from the 8th International Symposium
on Spatial Data Handling, ed. T.K. Poiker and N. Chrisman.
Burnaby, BC: IGU. 308–20.
———.
‘‘Geographical
Categories:
Investigation.’’ International Journal
Information Science 15: 591–612.
An
Ontological
of Geographical
Smith, N. 1992. ‘‘History and Philosophy of Geography:
Real Wars, Theory Wars.’’ Progress in Human Geography 16:
257–71.
cartographica (volume 40, issue 4)
Please
update
page range.
[Ver: A3B2 8.07r/W] File: {UTPress}Cartographica/CARTO-40(4)-005.3d
Creator: / Date/Time: 8.11.2005/10:38am Page: 47/62
Social Perspectives on Semantic Interoperability: Constraints on Geographical Knowledge from a Data Perspective
Stock, K., and D. Pullar. 1999. ‘‘Identifying Semantically
Similar Elements in Heterogeneous Spatial Databases Using
Predicate Logic Expressions.’’ In Interoperating Geographic
Information Systems, ed. A. Vckovski, K.E. Brassel, and
H.J. Schek. Zurich: Springer. 231–52.
Tosta, N. 1997. ‘‘National Spatial Data Infrastructures and the
Roles of National Mapping Organizations.’’ In Framework for
the World, ed. D. Rhind. Cambridge, UK: GeoInformation
International. 173–86.
Towers, G. 2000. ‘‘Applying the Political Geography of Scale:
Grassroots Strategies and Environmental Justice.’’
Professional Geographer 52: 23–36.
Vckovski, A., K.E. Brassel, and H.-J. Schek, eds. 1999.
Interoperating Geographic Information Systems. New York:
Springer.
Voisard, A., and H. Schweppe. 1998. ‘‘Abstraction and
Decomposition in Interoperable GIS.’’ International Journal
of Geographical Information Science 12: 315–33.
Volta, G.S., and M.J. Egenhofer. 1993. ‘‘Interaction with GIS
Attribute Data Based on Categorical Coverages.’’ In Spatial
Information Theory: A Theoretical Basis for GIS, COSIT ’93,
ed. A.U. Frank and I. Campari. Berlin: Springer. 15–31.
Walton, J. 1995. ‘‘How Real(ist) Can You Get?’’ Professional
Geographer 47: 61–65.
Ungerer, M.J., and M.F. Goodchild. 2002. ‘‘Integrating Spatial
Data Analysis and GIS: A New Implementation Using the
Component Object Method (COM).’’ International Journal of
Geographical Information Science 16/11: 41–53.
Waterton, C. 2002. ‘‘From Field to Fantasy: Classifying
Nature, Constructing Europe.’’ Social Studies of Science 32:
177–204.
Vckovski, A. 1999a. ‘‘Interoperability and Spatial Information
Theory.’’ In Interoperating Geographic Information Systems,
ed. M.F. Goodchild, M. Egenhofer, R. Fegeas, and C. Kottman.
Boston: Kluwer Academic. 31–37.
Weibel, R., and G. Dutton. 1998. ‘‘ConstraintBased Automated Map Generalization.’’ In Proceedings
from the 8th International Symposium on Spatial Data
Handling, ed. T.K. Poiker and N. Chrisman. Burnaby, BC: IGU.
214–24.
———. 1999b. ‘‘Interoperability by Exchanging Executable
Content.’’ In Interoperating Geographic Information
Systems, ed. M.F. Goodchild, M. Egenhofer, R. Fegeas, and
C. Kottman. Boston: Kluwer Academic. 99–106.
cartographica (volume 40, issue 4)
Winter, S. 2001. ‘‘Ontology: Buzzword or Paradigm Shift in GI
Science?’’ International Journal of Geographical Information
Science 15: 587–90.
61
[Ver: A3B2 8.07r/W] File: {UTPress}Cartographica/CARTO-40(4)-005.3d
Creator: / Date/Time: 8.11.2005/10:38am Page: 47/62
cartographica
the international journal for geographic information and geovisualization