“When I use a word,” Humpty Dumpty said, in a
rather scornful tone, “it means just what I choose
it to mean, neither more nor less.”
From Through the Looking Glass, L. Carroll
Visualizers, like logicians, have long been concernedwith meaning. Generalizing from MacEachren’s
overview of cartography,1 visualizers have to think
about how people extract meaning from pictures (psy-
chophysics), what people understand from a picture
(cognition), how pictures are imbued with meaning
(semiotics), and how in some cases that meaning aris-
es within a social and/or cultural context. If we think
of the communication acts carried out in the visualiza-
tion process (see Figure 1), further levels of meaning
are suggested. In the figure, visualization begins when
someone has data that they wish to explore and inter-
pret; the data are encoded as input to a visualization
system, which may in its turn interact with other sys-
tems to produce a representation. This is communicat-
ed back to the user(s), who have to assess this against
their goals and knowledge, possibly leading to further
cycles of activity.
Each phase of this process involves communication
between two parties. For this to succeed, those parties
must share a common language with an agreed mean-
ing. For example, when someone passes a data set to a
visualization tool, it is with some understanding of how
the tool will interpret content of the data set, and how
to interpret the output of algorithms that the tool might
apply to the data. This agreed meaning can arise and be
expressed in many different ways. We offer the follow-
ing three steps, in increasing order of formality:
1. terminology (jargon),
2. taxonomy (vocabulary), and
3. ontology.
The terminology level introduces the meaning of con-
cepts and expresses them informally, through, for exam-
ple, a glossary or published papers. The organization of
concepts is ad hoc and not in itself machine processable.
This step includes concepts where the concept itself
might be given a precise mathematical definition;
although the definition is precise within the body of the-
ory in which it is located, shared meaning of the con-
cept relies on social and cultural mechanisms.
In the taxonomy level, a definition of concepts
remains informal, but the concepts themselves are orga-
nized in some structured way. The organization of the
concept provides some context in which concepts can
be related and compared. However, because the orga-
nization itself need not follow any particular set of rules,
the taxonomy is not machine processable and opera-
tions on multiple taxonomies (for example, compari-
son, union, and so on) require understanding and
interpretation of the basis on which the taxa are formed.
The ontology level describes concepts using a set of
constructors with a preagreed meaning, for example,
through a set of relationships that can be asserted
between primitives. Because there is a fixed way of
defining new concepts, it’s possible for an ontology to
be made machine processable. This extends to opera-
tions across multiple ontologies.
To date, much of the knowledge about visualization
data, processes, and representations is at level 1 (termi-
nology)—for example, in the definition of data sets, doc-
umentation of procedural interfaces, and theories from
cognate disciplines. However, there has been work to
organize this knowledge, resulting in a number of tax-
onomies and models that formalize aspects of the visu-
alization process (at level 2, or taxonomy). Our
argument in this article is that it’s time to begin synthe-
sizing these fragments and views into a level 3 model, an
ontology of visualization. We also address why this
should happen, what is already in place, how such an
ontology might be constructed, and why now.
Motivation
We give four reasons for seeking a more rigorous
foundation for visualization:
■ collaboration,
■ composition,
■ preservation (curation), and
■ education.
We now expand each of these points.
Visualization is a collaborative activity involving
domain and system experts, and sometimes multiple
visualization systems. A shared vocabulary might be suf-
ficient for human-to-human collaboration, but if we
want to support remote collaboration via software tools,
a greater level of formalization is required.
Developments in the Web services, Semantic Web, and
Grid communities, of which we will have more to say
David J. Duke
and Ken .W.
Brodlie
University of
Leeds
David .A. Duce
Oxford Brookes
University
Ivan Herman
Centrum voor
Wiskunde en
Informatica
(CWI)
Do You See What I Mean?____________________________
Visualization Viewpoints
Editor: Theresa-Marie Rhyne
2 March/April 2005 Published by the IEEE Computer Society 0272-1716/05/$20.00 © 2005 IEEE