Вы находитесь на странице: 1из 12

What Makes a Good Diagram?

Improving the
Cognitive Effectiveness of Diagrams in IS Development
Daniel Moody
Department of Information Systems and Change Management
University of Twente, Enschede, Netherlands
E-mail: d.l.moody@bbt.utwente.nl

Abstract. Diagrams play a critical role in IS development. Despite this, ISD practitioners
receive little or no instruction on how to produce good diagrams. In the absence of this,
they are forced to rely on their intuition and experience, and make layout decisions that
distort information or convey unintended meanings. The design of ISD graphical notations
is ad hoc and unscientific: choice of conventions is based on personal taste rather than sci-
entific evidence. Also, existing notations use a very limited graphic vocabulary and thus
fail to exploit the potential communication power of diagrams. This paper describes a set of
principles for producing good diagrams, which are defined as diagrams that communicate
effectively. These provide practical guidance for both designers and users of ISD diagram-
ming notations and are soundly based on theoretical and empirical evidence from a wide
range of disciplines. We conclude that radical change is required to ISD diagramming prac-
tices to achieve effective user-developer communication.
Diagrams form a critical part of the language of IS development: most ISD
techniques rely heavily on graphical representations. For example, UML 2.0 con-
sists of 13 types of models, all of which are represented in graphical form [29].
The primary reason for using diagrams in IS development is to facilitate commu-
nication. In particular, diagrams are believed to be more effective than text for
communicating with end users. Effective user-developer communication is critical
for successful development of information systems.
1.1 Cognitive Effectiveness: What is a Good Diagram?
A diagram is a sentence in a graphical language [24]. The primary purpose of any
language is to communicate. Therefore a good diagram is one which communi-
cates effectively. Communication (or cognitive) effectiveness is measured by the
speed, ease and accuracy with which the information content can be understood.
The cognitive effectiveness of diagrams is one of the most widely held assump-
tions in the ISD field. However cognitive effectiveness is not an intrinsic property
of diagrams but something that must be designed into them [19].
1.2 Current State of Practice
Despite the importance of diagrams in IS development, practitioners typically
receive little or no instruction in how to produce effective diagrams. As a result,
they are forced to rely on their intuition and experience (which is often wrong),
and make layout decisions that distorts information or conveys unintended
1. INTRODUCTION
judgements about aesthetics (what looks good), which they are not qualified to
make as they typically lack expertise in graphic design. Also, what looks good is
not always what communicates most effectively. Current ISD diagramming prac-
tices are based on erroneous assumptions about what makes diagrams effective
and flawed heuristics about how best to construct them. Examples of commonly-
used heuristics include avoiding line crossings, using zigzag lines and expanding
symbols to fit labels. While these are designed to improve readability of diagrams,
in most cases they have the opposite effect. Despite this, such practices have been
perpetuated over time and have become so entrenched that they are often docu-
mented as best practices [e.g. 1].
The design of ISD diagramming notations is also largely ad hoc and unscien-
tific. Decisions about how to graphically represent constructs are based on
personal taste, intuition or consensus rather than on scientific evidence. There is
usually no theoretical or empirical justification for conventions used, perhaps re-
flecting a belief that it does not really matter which conventions are chosen. ISD
diagramming notations also use a perceptually limited repertoire of graphical tech-
niques and thus fail to exploit the potential power of diagrams [33]. The same
graphical symbols (variants of boxes and arrows) are used over and over again
while some of the most effective graphical techniques such as colour, spatial lay-
out, size and value are not used at all or used informally [33]. Finally, most ISD
diagramming notations are inconsistent with principles of graphic design. This is
not surprising as designers of ISD notations typically lack training or expertise in
graphic design. However while it is common in other areas of ISD practice (e.g.
user interface design, web development) to get advice from graphic design spe-
cialists, notation designers rarely do the same.
1.3 Current State of Research
The perceptual characteristics (visual appearance or form) of diagrams have been
grossly understated by ISD researchers. While issues of semantics or content
(what constructs to include in a notation) are treated as matters of substance, de-
tails of graphical syntax (how to visually represent these constructs) are treated as
being of little or no consequence. Choice of graphical conventions is seen by re-
searchers (like notation designers) as a matter of aesthetics or personal taste rather
than effectiveness [14]. Research in diagrammatical reasoning suggests the exact
opposite: the cognitive effectiveness of diagrams is primarily determined by their
perceptual characteristics [19, 33]. Even slight changes in graphical representation
can have dramatic impacts on understanding and problem solving performance.
The extent to which diagrams exploit perceptual features largely explains why
some diagrams are effective and others are not [19]. This suggests that the form of
diagrams is just as important if not more than their content.
1.4 Objectives of this Paper
Most ISD diagrams do not communicate effectively and actually act as a barrier
rather than an aid to user-developer communication [17, 28, 39]. Field studies
Daniel Moody
messages. Decisions about presentation of diagrams tend to be driven by subjective
482
show that end users understand ISD diagrams very poorly and that most develop-
ers dont even show diagrams to their users [13, 14, 28, 37]. The fault for this lies
with educators, notation designers and researchers, who have largely ignored is-
sues of graphical representation or treated them in an ad hoc way. The aim of this
paper is to provide a scientific basis for the use of diagrams in IS development.
We argue that ISD diagramming practice should be evidence based: decisions
about what graphical conventions to use (language level) and layout of individual
diagrams (sentence level) should be based on evidence about cognitive effective-
ness rather than subjective notions of aesthetics. The major focus of this research
is on improving the effectiveness of communication with non-specialists (end
users) as this is most critical for improving for improving the quality of the IS
development process.
In order to produce more cognitively effective diagrams, we need to consider two
things [49]:
The language of graphics: the techniques available for encoding information
graphically. Clearly, the better our command of the language, the more ef-
fectively we will be able to communicate.
Human graphical information processing: how diagrams are processed by
the human mind. This is necessary to evaluate the cognitive effectiveness of
alternative representations.
2.1 The Language of Graphics
The seminal work in the field of graphical communication is Jacques Bertins
Semiology of Graphics [5]. This is considered by many to be for graphic design
what the periodic table is for chemistry. Bertin identified eight elementary visual
variables which can be used to graphically encode information (Figure 1). These
are categorised into planar variables (the two spatial dimensions) and retinal vari-
ables (features of the retinal image).

Figure 1. Visual Variables [5]
What Makes a Good Diagram? 483
2. THEORIES OF GRAPHICAL
COMMUNICATION
set of atomic building blocks that can be used to construct any graphical represen-
tation similar to the way the elements of the periodic table can be used to
construct any chemical compound. Different visual variables are suitable for en-
coding different types of information. For example, colour can be used to encode
nominal data but not ordinal or ratio data because it is not psychologically ordered
[18]. The choice of visual variables has a major impact on cognitive effectiveness
as it affects both speed and accuracy of interpretation [9, 22, 50].
2.2 Human Graphical Information Processing
Figure 2 shows a model of human graphical information processing, which re-
flects current research in human cognition and visual perception. This consists of
four major processing stages:
Perceptual discrimination: features of the retinal image (i.e. visual variables)
integration) [21, 41]. Based on this initial processing, the diagram is parsed
into discrete elements based on common visual properties and separated
from the background (figure-ground segregation) [30, 50].
Perceptual configuration: diagram elements are grouped together (possibly
recursively) based on their visual characteristics [30, 50]. The Gestalt Laws
of Perception define a set of rules for how visual stimuli are organised into
perceptual groups [46].
Cognitive Processing Perceptual Processing
Perceptual
discrimination
Perceptual
configuration
Working
memory
Long term
memory
attention
Diagram
retinal
image

Figure 2. Model of Graphical Information Processing
Working memory (WM): all or part of the processed image is brought into
WM for active processing and interpretation under conscious control of at-
tention. Perceptual precedence determines the order in which elements are
attended to [51]. WM is a temporary storage area which synchronises rapid
perceptual processes with slower cognitive processes. It has very limited ca-
pacity and duration and is a known bottleneck in graphical information
processing [18, 22].
Long term memory (LTM): information extracted from the diagram is inte-
grated with prior knowledge stored in LTM. LTM is a permanent storage
area which has unlimited capacity and duration but is relatively slow. There
are two types of prior knowledge relevant to diagram understanding: domain
knowledge (knowledge about the represented domain) and notational knowl-
edge (knowledge about the diagramming notation). In the case of notation
experts (i.e. IS developers), notational knowledge is likely to be encoded in a
diagram schema, which largely automates the process of diagram interpreta-
Daniel Moody
are detected, some serially, some in parallel, across the visual field (feature
The set of visual variables define a vocabulary for graphical communication: a
484
tion [34]. Domain experts (i.e. end users) are unlikely to have such a schema
as they interact with such diagrams infrequently and therefore interpretation
will be much more error-prone and will require significant conscious effort.
Perceptual processes are automatic, pre-attentive, very fast and mostly executed in
parallel while cognitive processes operate under conscious control of attention and
are relatively slow, effortful and sequential. A major basis for the cognitive advan-
tages of diagrams is that they shift some of the processing burden to the
perceptual system, freeing up cognitive resources for other tasks [36].
In this section, we define a set of evidence-based principles for producing cogni-
tively effective ISD diagrams. These are based on theoretical and empirical
evidence from a wide range of disciplines including cartography, conceptual mod-
elling, cognitive psychology, communication theory, computer graphics,
diagrammatic reasoning, education, graphic design, graph drawing, human-
computer interaction, information visualisation, linguistics, multimedia design,
psychophysics, semiotics, statistics, technical writing, typography, visual percep-
tion and visual programming.
3.1
Perceptual discrimination is the first step in graphical information processing
(Figure 2). Accurate discrimination of diagram elements is a necessary prerequi-
site for further processing. There are two aspects to discriminability [18]:
Absolute discriminability: the ability to see diagram elements and separate
them from the background.
Relative discriminability: the ability to differentiate between different types
of diagram elements.
Absolute discriminability is determined by three primary factors:
Size: diagram elements (and also textual labels) need to be a certain mini-
mum size to be seen and recognised correctly [48]. The optimal size of
elements for human perception have been empirically established for both
visual elements and text [2, 40].
Contrast: according to the Gestalt Figure-Ground principle, the greater the
contrast between diagram elements and the background, the more readily ob-
jects will be detected and recognised [30, 46]. Contrast can be achieved by
using clearly different surface properties (colour, texture or value) between
diagram elements and the background. Adequate contrast should also be es-
tablished between textual labels and their background.
Proximity: discernibility of diagram elements decreases with proximity of
other elements [52]. This relates to the use of white space, which is one of
What Makes a Good Diagram? 485
3. PRINCIPLES FOR PRODUCING
EFFECTIVE DIAGRAMS
Discriminability: Diagram Elements should be Easy
to See and Differentiate from one another
the major design elements in graphic design [47]. There is no empirical basis
for choosing the optimal spacing between elements but for most types of ISD
diagrams 11.5 shape widths provides adequate separation.
Relative discriminability is determined by the number and size of differences be-
tween symbols used to represent different constructs. The greater the perceptual
variation between symbols used to represent different constructs, the faster and
more accurately they will be recognised [51]. If differences are too subtle, errors
in interpretation and ambiguity can result. In particular, requirements for dis-
criminability are much higher for novices (end users) than for experts [7, 8].
3.2 Manageable Complexity: Diagrams should not
One of the most common mistakes in ISD diagramming practice is to show too
much information on a single diagram. This results in absurdly complex dia-
grams that are a barrier rather than an aid to communication [17]. The reason for
this is that the amount of information that can be effectively conveyed by a single
diagram is limited by human perceptual and cognitive abilities:
Perceptual limits: The ability to visually discriminate between diagram ele-
ments decreases with their number and proximity [23]. In general, the
difficulty of discerning diagram elements increases quadratically with dia-
gram size [31]. The root cause of discriminability problems with ISD
diagrams (Principle 1) is excessive complexity.
Cognitive limits: The number of diagram elements that can be comprehended
at a time is limited by working memory capacity, which is believed to be
seven plus or minus two concepts at a time [3, 21, 45]. When this is ex-
ceeded, a state of cognitive overload ensues and comprehension degrades
rapidly [20, 26].
One of the most effective ways of reducing the complexity of large systems is to
divide them into smaller subsystems or modules. This is called decomposition or
modularisation [4]. In the context of diagrams, this means that large diagrams
should be divided into perceptually and cognitively manageable chunks (seven
plus or minus two elements per diagram). Experimental studies show that modu-
larising ISD diagrams in this way improves end user understanding and
verification accuracy by more than 50% [27].
3.3
In most ISD diagrams, all elements look the same: there is no way of telling which
are most important [17]. Such representations act as very poor information filters.
Also, because there is no clear entry point or processing sequence, it makes them
hard to access for novices and leads to inefficient and haphazard processing [6,
33]. The visual variables (Figure 2) can be used to create a clear perceptual prece-
dence or visual hierarchy among diagram elements. The most effective visual
variables for emphasis are those suitable for encoding ordinal data as emphasis
Daniel Moody
Exceed Perceptual and Cognitive Limits
Emphasis: The Relative Importance of Diagram
Elements should be Clearly Shown
486
defines a partial ordering over diagram elements. The most important concepts
should be emphasised (highlighted) to bring them to the readers attention, while
less important or background elements should be de-emphasised (lowlighted)
[52]. This defines a clear processing sequence which facilitates more efficient
processing [6, 51]. Research in diagrammatic reasoning shows that directing at-
tention to the most important concepts can dramatically improve understanding
and problem-solving performance [6, 11, 52].
3.4
One of the unique problems with ISD diagrams compared to diagrams used in
most other disciplines is that systems are typically represented using multiple dia-
grams. For example, UML consists of 13 different types of diagrams, each of
which represents the system from a different perspective. The need to manage
complexity (Principle #2) further exacerbates the problem by creating multiple
diagrams of each type. This results in a complex network of diagrams that can be
difficult to understand as a whole and navigate through. Using multiple diagrams
places additional cognitive demands on the user to mentally integrate information
from different diagrams and to keep track of where they are in the network of dia-
grams [38, 42]. For such representations to be cognitively effective, the
diagramming notation must provide explicit mechanisms to support [16]:
Conceptual integration: enabling the reader to integrate information from
separate diagrams into a coherent mental representation of the problem.
Perceptual integration: providing perceptual cues (orienting, contextual and
directional information) to aid navigation between diagrams.
There are a range of mechanisms which can be used to achieve cognitive integra-
tion, such as summary diagrams [16], signposting [25] and spatial contiguity [44].
3.5
Perceptually direct representations are representations whose interpretation is
spontaneous or natural, in that their meaning can be extracted automatically by the
perceptual system. Such representations are highly efficient as they offload inter-
pretation effort from the cognitive system to the perceptual system: extracting
meaning from the diagram becomes effort-free.
Representation of constructs: Icons are symbols which perceptually resemble
the objects they represent [32]. Using icons to represent constructs reduces
cognitive load because they have built-in mnemonics: the association with
the referent concept can be perceived directly, and does not have to be learnt
[33]. Icons also make diagrams more visually appealing, speed up recogni-
tion and recall, and improve intelligibility to nave users [7, 8].
Representation of relationships: Perceptual directness also applies to repre-
sentation of relationships among diagram elements. Certain spatial
What Makes a Good Diagram?
Cognitive Integration: When Multiple Diagrams are
used, Explicit Mechanisms should be Included to
Support Cognitive Integration
Perceptual Directness: Make use of Perceptually
Direct Representations
487
configurations of diagram elements predispose people towards a particular
interpretation even before the meaning of the elements is understood. For
example, left to right arrangement of objects suggests causality or sequence
while placing objects inside other objects suggests class membership [12,
48]. Diagramming notations can be designed to exploit these intuitively un-
derstood spatial arrangements to increase ease and accuracy of interpretation.
3.6
In most ISD diagrams, there is no clear structure or grouping among diagram ele-
ments. This leads to inefficient encoding in WM, as each diagram element must be
encoded as a separate chunk [18]. The Gestalt Laws of Perception define a set
of empirically validated principles for how the human perceptual system organises
visual stimuli into perceptual units [46]. These can be used to group related dia-
gram elements together and so facilitate perceptual configuration (Figure 3). In
this way, This facilitates more efficient use of WM, as each group of elements can
be encoded as a chunk rather than encoding each element separately [26]. This
reduces cognitive load and frees up cognitive resources for other information
processing tasks [10, 22]. Perceptual grouping provides both an alternative and a
complement to decomposition (Principle 2). Organising diagram elements into
groups expands the number of elements that can be shown on each diagram with-
out exceeding cognitive limits [10]. This reduces the total number of diagrams
required, which in turn reduces the need for cognitive integration (Principle 4).
3.7
Identification is an important concept in cartography [35] but is given little ex-
plicit attention in ISD diagramming practice. There are two aspects to
identification [5]:
External identification: this defines the correspondence between the diagram
and the represented world. Each diagram should have a title, which should
be clearly recognisable as such by its size and placement [18]. This should
summarise the content of the diagram (the part of the referent domain it rep-
resents). Diagram elements (both nodes and links) should also be clearly
labelled, using terminology familiar to domain experts to help trigger domain
knowledge in LTM. Labels should be clearly grouped with their referent ob-
jects using Gestalt principles [15].
Internal identification: this defines the correspondence between graphical
conventions and their meaning. The diagram type should be clearly identi-
fied to trigger the appropriate diagram schema in LTM (if one exists) and
reduce likelihood of misinterpretation. In addition, a legend or key should be
included, summarising the graphical conventions used. This should be in-
cluded within the frame of each diagram rather than on a separate sheet of
paper or document to avoid problems of cognitive integration [25, 43, 44].
Daniel Moody
Structure: Organise Diagram Elements into
Perceptual Groups
Identification: Diagrams should be Clearly Labelled
488
3.8
Most ISD diagramming notations use a very limited graphical vocabulary and use
only one of the eight available visual variables to encode information: shape.
Shape is one of the least cognitively efficient variables as it can only be used to
encode nominal data and is detected serially [23]. Visual expressiveness refers to
the number of different visual variables used to encode information. Using multi-
ple visual variables results in a perceptually enriched representation which uses
multiple, parallel channels of communication. It also helps increase visual interest
and attention. Using multiple visual variables to convey the same information
(redundant coding) improves accuracy of communication and counteracts noise.
3.9
Graphic complexity is defined as the number of different graphical conventions
used in a notation [28]. Experimental studies show that human ability to discrimi-
nate between perceptually distinct alternatives (the span of absolute judgement) is
around six categories [26]. Most ISD diagramming notations currently in use ex-
ceed this significantly: for example, UML Class Diagrams have a graphic
complexity of 14. One solution to this problem is increase the number of percep-
tual dimensions (i.e. visual variables) on which constructs differ. This has been
shown empirically to increase human ability to discriminate between stimuli [26].
Another solution is not to show everything in graphical form: diagrams are useful
for showing some types of information (e.g. structure, relationships) but not oth-
ers (e.g. detailed business rules): some information can be more effectively
represented in textual form [33].
This paper has described a set of principles for producing good diagrams, which
are defined as diagrams that are cognitively effective (i.e. that communicate effec-
tively). These principles are soundly based on theoretical and empirical evidence
from a range of disciplines rather than on intuition, experience and convention
like most principles used in ISD practice. The principles can be either applied at
the level of diagramming notations (language level) or individual diagrams (sen-
tence level). Some principles apply mainly at the language level, others apply
mainly at the diagram level but most apply at both to at least some extent. In ap-
plying these principles to current ISD diagramming practices, the conclusion is
that radical change is required to achieve effective user-developer communication.
4.1 Theoretical Significance
The theoretical contributions of this paper are as follows:
(a) It highlights the importance of the visual aspects (form) of ISD notations,
which have been grossly understated in ISD research and practice.
What Makes a Good Diagram?
Visual Expressiveness: Use the Fu ll Range of Visual
Variables
Graphic Simplicity: The Numbe r o f Different
Graphical Conventions should be Limited
489
4. CONCLUSION
(b) It defines a descriptive model of how diagrams communicate, based on re-
search in graphic design, cognition and visual perception. This provides a
theoretical basis for evaluating the visual aspects of ISD notations which
complements methods used to evaluate semantic aspects.
(c) It defines a set of prescriptive principles for producing cognitively effec-
tive diagrams. These define causal relationships between the perceptual
characteristics of diagrams and efficiency and effectiveness of human in-
formation processing. These help to increase our understanding of what
makes ISD diagrams effective or ineffective and also represent theoretical
propositions which can be empirically tested.
4.2 Practical Significance
The principles defined in this paper provide practical guidance for both designers
and users of ISD diagramming notations. Importantly, they are not abstract, theo-
retical principles but highly specific and operational principles, which could be
easily incorporated into current ISD practices. They can be used by ISD practitio-
ners to produce diagrams that communicate more effectively with their customers.
They can also be used by ISD notation designers to develop diagramming nota-
tions that more effectively exploit human perceptual and cognitive capabilities.
3.
4.
6.
7.
8.
9.
10.
1980. 208: p. 1181-1182.
Daniel Moody
1. Ambler, S.W., The Elements of UML 2.0 Style. 2005, Cambridge, England: Cambridge
University Press.
Psychological Research, 1984. 46(247-259).
2. Antes, J.R. and S.W. Mann, Global-Local Precedence in Picture Processing.
Baddeley, A. and G. Hitch, Working Memory, in The Psychology of Learning and
Motivation: Volume 8, G.H. Bower, Editor. 1974, Academic Press: London.
Baldwin, C.Y. and K.B. Clark, Design Rules Volume 1: The Power of Modularity.
2000, Cambridge, Massachuesetts, USA: MIT Press.
5. Bertin, J., Semiology of Graphics: Diagrams, Networks, Maps. 1983, Madison ,
Wisconsin, USA: University of Wisconsin Press.
Blankenship, J. and D.F. Dansereau, The Effect of Animated Node-Link Displays on
Information Recall. The Journal of Experimental Education, 2000. 68(4): p. 293-308.
Britton, C. and S. Jones, The Untrained Eye: How Languages for Software Specification
Support Understanding by Untrained Users. Human Computer Interaction, 1999. 14:
p. 191-244.
Britton, C., S. Jones, M. Kutar, M. Loomes, and B. Robinson. Evaluating the Intelligibility
of Diagrammatic Languages Used in the Specification of Software. in Proceedings of
of the First International Conference on the Theory and Application of Diagrams
(Diagrams 2000). 2000. Edinburgh, Scotland.
Cleveland, W.S. and R. McGill, Graphical Perception: Theory, Experimentation and
Application to the Development of Graphical Methods. Journal of the American
Statistician Association, 1984. 79(387): p. 531-554.
Ericsson, K.A., W.G. Chase, and S. Faloon, Acquisition of a Memory Skill. Science,
490
REFERENCES
11.
The Entity Relationship Model. European Journal of Information Systems, 1995. 4(1):
p. 31-40.
2002. 9(10).
15. Imhof, E., Positioning Names on Maps. The American Cartographer, 1975. 2: p. 128-
144.
16.
17. Kimball, R., Is ER Modeling Hazardous to DSS? DBMS Magazine, 1995.
1989. 3: p. 185-226.
19.
Words. Cognitive Science, 1987. 11(1).
23.
25.
Capacity For Processing Information. The Psychological Review, 1956. 63: p. 81-97.
28.
30.
Sapienza, Italy.
What Makes a Good Diagram?
Grant, E.R. and M.J. Spivey. Guiding Attention Produces Inferences in Diagram-
Based Problem Solving. in First International Conference on the Theory and Application
Application of Diagrams (Diagrams 2000). 2000. Edinburgh, Scotland.
12. Gurr, C.A., Effective Diagrammatic Communication: Syntactic, Semantic and Pragmatic
Issues. Journal of Visual Languages and Computing, 1999. 10: p. 317-342.
13. Hitchman, S., Practitioner Perceptions On The Use Of Some Semantic Concepts In
14. Hitchman, S., The Details of Conceptual Modelling Notations are Important - A
Comparison of Relationship Normative Language. Communications of the AIS,
Kim, J., J. Hahn, and H. Hahn, How Do We Understand a System with (So) Many
Systems Research, 2000. 11(3): p. 284-303.
Diagrams? Cognitive Integration Processes in Diagrammatic Reasoning.Information
18. Kosslyn, S.M., Understanding Charts And Graphs. Applied Cognitive Psychology,
Larkin, J.H. and H.A. Simon, Why a Diagram is (Sometimes) Worth Ten Thousand
21. Lohse, G.L., A Cognitive Model for Understanding Graphical Perception. Human-
22. Lohse, G.L., The Role of Working Memory on Graphical Information Processing.
Behaviour and Information Technology, 1997. 16(6): p. 297-308.
Computer Interaction, 1993. 8(4): p. 353-388.
hensive Psychiatry, 1975. 16(3): p. 105-124.
20. Lipowski, Z.J., Sensory And Information Inputs Overload: Behavioural Effects.Compre-
Lohse, G.L., D. Min, and J.R. Olson, Cognitive Evaluation of System Representation
Diagrams. Information & Management, 1995. 29: p. 79-94.
24. Mackinlay, J., Automating the Design of Graphical Presentations of Relational
Mayer, R.E. and R. Moreno, Nine Ways to Reduce Cognitive Load in Multimedia
Learning. Educational Psychologist, 2003. 38(1): p. 43-52.
26. Miller, G.A., The Magical Number Seven, Plus Or Minus Two: Some Limits On Our
27. Moody, D.L. Complexity Effects On End User Understanding Of Data Models: An
Experimental Comparison Of Large Data Model Representation Methods. in
2002. Gdansk, Poland.
Nordbotten, J.C. and M.E. Crosby, The Effect of Graphic Style on Data Model
Interpretation. Information Systems Journal, 1999. 9(2): p. 139-156.
29. OMG, Unified Modeling Language Version 2.0: Superstructure. 2005: Object
Management Group (OMG).
Palmer, S. and I. Rock, Rethinking Perceptual Organisation: The Role of Uniform
Connectedness. Psychonomic Bulletin and Review, 1994. 1(1): p. 29-55.
31. Patrignani, M., Visualization of Large Graphs, in Dottorato di Ricerca (Doctoral
Dissertation), Ingegneria Informatica. 2003, Universit degli Studi di Roma: La
32. Peirce, C.S., Charles S. Peirce: The Essential Writings (Great Books in Philosophy),
ed. E.C. Moore. 1998, Amherst, USA: Prometheus Books.
Information. ACM Transactions on Graphics, 1986. 5(2): p. 110-141.
Proceedings of the Tenth European Conference on Information Systems (ECIS2002).
491
35.
36.
39. Tasker, D., Worth 1,000 Words? Ha! Business Rules Journal, 2002. 3(11).
41.
Psychology, 1980. 12: p. 97-136.
42.
303-317.
43.
44.
Type. 2002, New York: Allworth Press.
52.


Daniel Moody
Programming. Communications of the ACM, 1995. 38(6): p. 33-44.
34. Pinker, S., A Theory of Graph Comprehension, in Artificial Intelligence and the Future
of Testing, R. Freedle, Editor. 1990, Lawrence Erlbaum & Associates: Hillsdale, New
Jersey, USA. p. 73-126.
Robinson, A.H., J.L. Morrison, P.C. Muehrcke, A.J. Kimerling, and S.C. Guptill, Elements
of Cartography (6th Edition). 1995, New York: Wiley.
Scaife, M. and Y. Rogers, External Cognition: How Do Graphical Representations
Work? International Journal of Human-Computer Studies, 1996. 45: p. 185-213.
37. Shanks, G.G., The Challenges Of Strategic Data Planning In Practice: An Interpretive
Case Study. Journal of Strategic Information Systems, 1997. 6(1): p. 69-90.
38. Siau, K., Informational and Computational Equivalence in Comparing Information
Modelling Methods. Journal of Database Management, 2004. 15(1): p. 73-86.

40. Tinker, M.A., Legibility of Print. 1963, Ames, Iowa, USA: Iowa State University Press.
Treisman, A. and G.A. Gelade, A Feature Integration Theory of Attention. Cognitive
Verdi, M.P., J.T. Johnson, W.A. Stock, R.W. Kulhavy, and P. Whitman-Ahern,
Organized Spatial Displays And Texts: Effects Of Presentation Order And Display
Type On Learning Outcomes. The Journal of Experimental Education, 1997. 65(4): p.
Wallgren, A., B. Wallgren, R. Persson, U. Jorner, and J.-A. Haaland, Graphing Statistics
& Data: Creating Better Charts. 1996, London: Sage Publications.
Watts-Perotti, J. and D.D. Woods, How Experienced Users Avoid Getting Lost in Large
Display Networks. International Journal of Human-Computer Interaction, 1999. 11(4):
169-299.
45. Weber, R.A., Are Attributes Entities? A Study Of Database Designers Memory
Structures. Information Systems Research, 1996. 7(2): p. 137-162.
46. Wertheimer, M., Laws of Organization in Perceptual Forms, in A Source Book of
Gestalt Psychology, W. Ellis, Editor. 1938, Routledge and Kegan Paul (originally
published in 1923 in German as Untersuchungen zur Lehre von der Gestalt II, in
Psycologische Forschung, 4, 301-350): London. p. 71-88.
47. White, A.W., The Elements of Graphic Design: Space, Unity, Page Architecture and
Transactions on Professional Communication, 1990. 33(3): p. 103-107.
48. Winn, W.D., Encoding and Retrieval of Information in Maps and Diagrams. IEEE
49. Winn, W.D., A Theoretical Framework for Research on Learning from Graphics.
International Journal of Educational Research, 1990. 14: p. 553-564.
50. Winn, W.D., Learning from Maps and Diagrams. Educational Psychology Review,
1991. 3: p. 211-247.
51. Winn, W.D., An Account of How Readers Search for Information in Diagrams.
Contemporary Educational Psychology, 1993. 18: p. 162-185.
Yeh, M. and C.D. Wickens, Attention Filtering in the Design of Electronic Map
Displays: A Comparison of Colour Coding, Intensity Coding and Decluttering
Techniques. Human Factors, 2001. 43(4): p. 543-562.
33. Petre, M., Why Looking Isnt Always Seeing: Readership Skills and Graphical
492

Вам также может понравиться