Minor Thesis Proposal

bonesworshipΤεχνίτη Νοημοσύνη και Ρομποτική

14 Νοε 2013 (πριν από 3 χρόνια και 5 μήνες)

54 εμφανίσεις

i



Minor

Thesis Proposal



Applying Graph Visualization to Represen
t Relational
Database
as a Quality Graph



By

Chaowen Huang

100110151

M
aster of Computer and Information Science



Supervisor

Dr. Jixue(Jerry) Liu



Due Date:
13
th

June 20
10




School of Computer and Information Science

University of South Australia

Mawson Lakes South Australi
a
ii


1.

Introduction

................................
................................
................................
.....

-

1
-

1.1

Background

................................
................................
................................
.

-

1
-

1.2 Motiva
tion

................................
................................
................................
...

-

2
-

1.3 Field of Thesis
................................
................................
.............................

-

3
-

1.4

Research Questions

................................
................................
.....................

-

3
-

1.5

Explication of Research Questions

................................
.............................

-

4
-

1.6

Scope and Limitation

................................
................................
..................

-

5
-

1.7

Contributions

................................
................................
...............................

-

5
-

2.

Literature review

................................
................................
............................

-

6
-

2.1 Relational Database Representation

................................
...........................

-

6
-

2.2 Graph Layout Types

................................
................................
...................

-

8
-

2.2.1 Original Graph Layout Types

................................
................................
........

-

9
-

2.2.1.1 Tree Layout (Classical Graph)

................................
....................

-

9
-

2.2.1.2 Sugiyama Layout (General Directed Graph)

............................

-

10
-

2.2.1.3 Grid Layout

................................
................................
...............

-

11
-

2.2.1.4 Spring Layout (Force
-
Directed Methods)

................................
.

-

11
-

2.2.1.5 Planar Layout

................................
................................
............

-

11
-

2.2.2 Modern Graph Layout Types

................................
................................
......

-

12
-

2.3 Graph Drawing Aesthetics

................................
................................
........

-

12
-

3.

Methodology

................................
................................
................................
..

-

14
-

3.1 Methods Adopted

................................
................................
......................

-

14
-

3.2 Research Process

................................
................................
.......................

-

14
-

3.3 Expected Outcome

................................
................................
....................

-

16
-

4.

Timetable

................................
................................
................................
.......

-

18
-

5.

Reference

................................
................................
................................
.......

-

19
-














-

1
-



1.

Introduction


1.1 Background


I
nformation visualization
which has been
becoming a more and more signific
ant field
in the last few years

discovers associations in the data set and visualizes data set with
association

[9]
.

G
raph visualiza
tion
which
is one of the subfields of the information
visualization

is different from the normal graph drawing

[10]
.

M
ore exactly, graph
visuali
zation not only

draws graphs

but also focuses on dealing with the graph
.

M
oreover, the application of graph visualization is more specific than the information
visualization.
B
y this
I

mean that graph visualization is only applied
to

the data set
which has

inherent relation
s

among data members

[10]
.


G
raph visualiza
tion has been widely applied t
o va
rious areas in recent years, for
example, data structures, entity
-
relationship diagrams (E
-
R diagram), data flow
diagrams,

semantic networks and knowledge
-
representation

diagrams, logic
programming,
document management systems, and
virtual reality

[10]
.

B
esides,
E
-
R
diagrams can be used for
representing database structure, especially relational
database

which makes the data member within the data set match each other by

certain relations

[4, 14]
.


The significance of database repre
sentation in database
study
ing and using has been
proved in the past survey
[1
3]
. More exactly, database representation affects user
study
ing and using database.
F
urthermore, different representations have different
-

2
-


influence on the user understanding and using database. By this I mean that different
representations may use diffe
rent semantics, symbols or methods of representing
relationship between data to represent the database. Even the same type of
representation has different drawing aesthetics.
A
dditionally, different drawing
aesthetics may have different impact on the user
und
erstanding of

graph

too

[18]
.



1.2 Motivation


N
ormally, in the relational database, relations are represented by relational tables
which contain their own attributes.
A
lthough

several table representation means have
been designed, including
table representation, table with arrows representation

and
tab
le with text only representation
, these table representation means

still have the
limitation

of representing the
entirety

information of the relational database.

F
or this
reason,
some representation methods
have been designed,
for example, logical data
str
ucture (LDS) representation,
these methods also have their own weaknesses.

F
or
example, by using LDS representation, understanding the exactly relationship
between attributes which belong to two relation tables are difficult for the users
.
B
ecause the representation of links is
showed
between tables
, and do not be showed
precisely

between the attributes in the tables

in LDS representation

[13]
.


A
ddition
ally, r
egular relational databases have, at least, hundreds of relation tables
and foreign keys (FKs) which indicate the relation between the relation tables.
For
this reason, it is difficult for use
r
s to study the relational database and get the useful
in
formation from the relational database

effectively
.


-

3
-


Based on the above discussion, it is obvious that d
atabase representation is significant
of database
study
ing and using.
M
oreover, relational
database
s have been widely used
in many other different
areas.

A
s a result,

the effectiveness of relational database
representation is b
ecoming more and more important

for relational database
.

1.3
Field of Thesis


D
atabase representation,

relational database representation,

graph visualization,
graph
layout,
an
d drawing aesthetic


1.4

Research Questions


T
he research questions of this study are as follows:



M
ain question:

How to d
raw a quality graph which can aid relational database users studying and
using relational database more effective
?



S
ub
-
questions:

1.

What
means of database representation should be used for

relational database
representation?

2.

W
hat types of
graph
layout should be used

for relational database
representation?

3.

W
hat drawing aesthetics should be used for designing and measuring the
relational
database representation?

4.

W
hat means of navigation and interaction should be used for relational
database representation?

-

4
-


1.5

Explication of Research Questions


T
here are many means of database representation, including E
-
R diagram
representation, logical data structure representation, table representation, table with
arrows representation, and table with text only representation
[13]
.
T
hey all have
their

own strengths and weaknesses.
F
or this reason, it is obvious when
designing

a
representation

method

should not only have the most numerous strengths among
previous representat
ion methods but also avo
id the most numerous weaknesses

in this
study.


A great many of
graph
layout
types

have been designed in the past as well, including
tree layout, grid layout, planar layout, etc
[10]
. They are all applied to different
applications of different drawing requirements. Thus, one of these layout types

would
be chosen to

modify so that
the graph layout
will

adapt

to this study
.


The five primary common drawing aesthetics have been validated by
Helen C.
Purchase

in 1997, for example,

minimizing

crosses and minimizing the number of
bends
[17]
. Two more
drawing aesthetics

have been
validated by
Helen C. Purchase

later, in 2002
[20]
.
T
hese common drawing aesthetics have different influence on user
understanding.
S
ince, not all these common drawing aesthetics are used in this study.


Several different navigation and interaction methods have been developed for aiding
users viewing the gra
ph, for example, zoom and pan technique.

O
ne of these
navigation and interaction methods would be chosen to use in this study.


-

5
-


1.6

Scope and Limitation


The primary purpose of this study is to modify the existing representation methods to
design

a quality rep
resentation method and the proper algorithm for this representation
method to represent the relational database

as a quality graph so that users can study
and use the
relational

database more effectively.

F
urthermore, this study may base on
basic drawing s
kills
,

the original graph layout algorithm (not the modern graph layout
algorithm)
, and the most effective
graph aesthetic to design the layout algorithm for
representing the relational database.

A
dditionally, the navigation and interaction
method is
chosen zoom and pan technique in this study.



T
hus, this study will focus on designing the layout algorithm of relational database
representation to draw a quality graph of relational database which can aid user
studying and using relational database more

effectively.


1.7

Contributions


T
he
re are several

contributions of this study

and they

are as follows:



This study will

be beneficial to

the relational database area, because this study
will
help

the relational model database

users

study and use

relational database
more effectively.



To some extent,

some
specific area
s

of the graph visualization
will be beneficial
from this study,
because
this study

will solve some similar problems which
also

exist

in those specific areas of the graph visualization.


-

6
-


2.

Literature review


The literatures

which

will
be
covered

in this section

are
all related to how to draw a
quality graph to represent the relational database. The related literatures are about the
previo
us database representation methods, graph layout types, graph drawing
aesthetics and navigation and interaction methods. Since the contents in this section
are
organized in
to

four

parts
:
section 2.1
-

will present

the related studies

of
relational
database
representation
;

the related studies

of various graph layout types

will be
demonstrated

in section 2.2
;

section 2.3

will
show

the related studies

of
graph
drawing aesthetics
.


2.1
Relational Database Representation


The relational database has been proposed

in 1970 by Edgar F. Codd

[5]
.
A
lthough
t
he study of relational database has been done many years
,

very
few

researches

have
been published
on relational database representation
.

T
here are a few
researches

on
the area of
database representation

and the effectiveness of different
database
representation.

These related researches will be showed in the following
part.


I
n the research
[15]
,
Lochovsky and Tsichritzis

design a m
ethod with a set of
measures to measure user performance in using database management systems
(DBMS).

F
or the part of

their study,

they compare
d

the effectiveness of three
different database representations, including hierarchical (tree) representation,
ne
twork representation, and relational (table) representation
.

Because

Lochovsky and
Tsichritzis

confused the database representation with query language features,

t
hey
got the results
which
were produced

by

using each representation in different
-

7
-


language
environment.

B
y this I mean that they could not confirm
whether

the results
are fully caused by the representation method.


C
omparing with
research
[15]
,

later in 1989,
the research
which
was published by

Kenny et al.

[12]

did the comparison of two

representation method, entity
-
relationship
representation and table representation,

in the same language environment, structured
query language

(SQL)
.

T
he result of this study is that there is no significant difference
of the two representation methods ap
plying to the same language environment, SQL.

But the result of this study which is the same as
[15]

could not be confirmed.

S
ince

Kenny et al.

could not confirm
whether

the results are fully caused by the
representation method
, like
[15]
.


T
he research

conducted by
Juhn and Naumann

[11]

is different from the researches
[15]

and
[12]
.

M
ore exactly,

Juhn and Naumann

directly
conducted the

comparison
effectiveness of four different database rep
resentations
, two semantic representations
and two table representations,

on user studying and using database

by
avoid
ing
using
a query language
.

T
he results of the study show that user understanding of
relationship by using semantic representation is
better than by using table
representation.

B
ut user
understanding of identifiers

by using table representation is
better than by using semantic representation.

One important discovery of this study is
that the table representation by adding arrows to repre
sent the relationship between
primary keys (PKs) and foreign keys (FKs) make user better understanding of
relationship of table representation.


T
o conclude, most previous researches have not focused on database representation
so
that
database representati
on
is
only a part of these researches. Some researchers of
previous researches may even confuse

the database representation with

query
-

8
-


language features.

T
here are little researches on database representation effectiveness
of user studying and using. It wa
s until 1996 that a research

on measuring the
effectiveness of database structure representation on database system studying and
using was

published by
Leitheise and March

[13]
.



2.2 Graph Layout

Types


G
raph layout types, in theory, are defined as graph layout

algorithm
s.
S
ince different
graph drawing requirements define the different graph
s
, which is the basic problem in
graph drawing too.

M
ore exactly,
the basic problem
can be described
as follows
:

supposed that there are a set of nodes and a set of edges (relations), where the nodes
should we draw and how to draw the edges (relations) between the nodes.

F
or this
reason
,
there is need to define different

types of graph and classify different layout
s

of
various graphs.

T
hat is, what graph layout types mean: what type of graph should use
what class of layout

[10]
.

A

majority of

the layout algorithms could be found in
previous researches organized by Battista et al.
[1]

or directly found in the graph
drawing book published by
Battista et al.

in 1998

[2]
.


A

great
many
of
studies
on

graph layout

have been published.

S
ince part of these
studies belong to graph drawing area and part of these studies belong to graph
visualization
area.

B
y this I mean that
some graph layout types are designed for
original graph drawing,

called original graph layout types,

but some other graph
layout types are designed for graph visualization drawing
, called modern graph layout
types
[10]
.

M
ore ex
actly,
in theory
,

original graph layout types are designed for
relative
ly

small graph drawing, while modern graph layout types are designed for
rel
atively large graph

drawing
.

F
or this reason,
the contents in this section are
organized in
to two

parts
: section 2.2.1
will
demonstrate

the related studies of various
-

9
-


original
graph layout types;

the related studies of various modern graph layout types

will be showed in section 2.2.2
.


2.2.1 Original Graph Layout Types


There are several original graph layout types

which were classified by Mutzel et al.
[16]
,
including
t
ree
l
ayout (
c
lassical
g
raph)
,
sugiyama layout (general directed graph),
grid layout, spring layout (Force
-
Directed graph), and planar layout.

T
he contents of
this section will be
showed in the following
five parts:

2.2.1.1 tree layout
(
c
lassical
g
raph)
; 2.2.1.2 sugiyama layout (general

directed graph); 2.2.1.3 grid layout; 2.2.1.4
spr
ing layout (Force
-
Directed Method
s
); and 2.2.1.5 planar layout.

,

2.2.1.1 Tree Layout (Classical Graph)


Tree layout, like its name, is the layout of an inverted tree.
Tree layout

is the most
well
-
known gr
aph layout type among all graph layout types

so that tree layout is the
classical representation and its layout graph is called classical graph.


The most famous layout algorithm in tree layout area may be the
Reingold and Tilford

[22]
.

Several other algorithms, like the
Reingold and Tilford
, are predictable
algorithms
, including
binary tree

[25]
, cone tree

[23]

and radial tree

[7]
.

M
ore exact
ly,
the predictable algorithm means that we can
predicate

the tree structure before
drawing.

B
ut it is obvious that we
cannot

predict the tree structure in every drawing.

F
or this reason, some tree algorithms are designed for solving this problem, but thes
e
new algorithms do not belong to original graph layout (tree layout) algorithm.
T
hus,
it
will not be mentioned and discussed
in this section.

-

10
-



T
he graphs drawn by
Reingold and Tilford

algorithm are classical and hierarchical
.

C
omparing with graphs drawn

b
y
Reingold and Tilford

algorithm,
graphs drawn by
radial tree algorithm will be more obscure for viewing the root of the tree so that
graphs drawn by
Reingold and Tilford

algorithm may be more hierarchical than
graphs drawn by radial tree algorithm.


2.2.1
.2 Sugiyama Layout (General Directed Graph)


T
he sugiyama layout was named by a researcher, Sugiyama, who conducted the study
on layout of general directed graph with the other researchers in
1981
[26]
.

N
ormally,
t
he
re are
three

steps in sugiyama layout algorithm

[16]
:




1.
Rank

a
ssignment



2.
Two layer crossing minimization



3.
H
ierarchy layout module

In the later study
[10]
,
Herman

added one more step in the
sugiyama layout algorithm
:
the final step, Subgraph (extraction)
.

B
ecause some other

studies

[17
-
19]

have prov
ed that minimizing crosses is significant
for improving the
readability

of graphs
,
the key step in the sugiyama layout algorithm
is the second step (Two layer crossing minimization).

T
he studies
[17
-
19]

will be
further discussed in the next section 2.3
-
graph drawing aesthetics.

T
he study
[8]

provided a research on minimizing crosses and the concept of minimizing crosses in
sugiyama layout algorithm is different.

M
ore exactly,
the concept of minimizing
crosses in sugiyama layout algorithm means minimiz
ing cro
sses between two
-

11
-


consecutive layers
, while
the concept of minimizing crosses in
[8]

means minimizing
cro
sses

in the whole graph.


2.2.1.3
Grid Layout


F
rom the study
[16]
,
grid layout algorithm has two main techniques:
Compaction
Module

and
Aug
mentation

Module, but both these two techniques
are not

important
for the graph visualization

for reviewing the these two techniques in
[2]
.

F
or this
reason, I do not discuss the
grid layout in this study.

I
f reader wants to get further
information on grid layout
, I recommend the book
[2]
.


2.2.1.4
Spr
ing Layout (Force
-
Directed Method
s
)


F
rom study
[16]
, t
he spring
layout

algorithm
has no certain techniques

for drawing

and also can be named as
Force
-
Directed Methods
.

T
he main purpose of
spring layout

is to make graph layout symmetry
, which was first
defined by
Eades in 1984
[6]
.

M
oreover, the effectiveness of symmetry
of the layout graph is excellent by using
spring layout algorithm.

In the later research in 1999,
Bertault
[3]

improved the spring
layout algorithm to make it predictable.



2.2.1.5 Planar Layout


B
ecause
the planar layout may be the constraint in

applications of
practical, t
he planar
layout is a significant topic in graph drawing
.

But in
graph visualization area,

it only
-

12
-


need measure the
planarity when drawing relatively small graph
[24]
.

N
ormally,
planar layout is not the mainly

topic in graph visualization

[2]
.


2.2.2 Modern Graph Layout Types


M
odern graph layout types are designed for relatively large graph drawing

and several
modern graph layout types have been designed in recent year
s

[10]
, including
3D
layout, spanning tree
s

layout
, hyperbolic Layout.

A
s mentioned above,

this study
would not base on the modern graph layout types.

T
hus,

I do not discuss the
modern
graph layout types in details

in this study


2.3 Graph

D
rawing

A
esthetics


The graph drawing has been studied for many years, but there are very little
researches on graph drawing aesthetics until 1995 that
three drawing aesthe
tics,
minimizing edge crossings, minimizing edge bends and maximizing symmetries, were

tested

by Helen C. Purchase et al.

[18]
.

M
oreover, two of the three
aesthetics
,
minimizing edge crossings and minimizing edge bends,

were validated
.


I
n 1997, a study conducted by Helen C. Purchase et al.
[21]

was pub
lished to
further
validate the graph drawing aesthetics.
T
his study
is nearly the same as
[18]

in 1995.


In the latter

1997
study
,
Helen C. Purchase

[17]

compared five common drawing
aesthetics and ranged them

as follows
:

-

13
-




1. Minimizing

c
rosses



2.

Minimizing
bends



3.

Maximizing
perceptual
s
ymmetry



4.

Maximizing

the o
rthogonality
structure



5.

M
aximising the minimum angles


A
fter 1997, the study in 2000
also by Helen C. Purchase

[19]
,
conduc
ted two
experiments to measure the effectiveness of aesthetics from user view point and
compare eight drawing algorithms from understanding view point
, respectively
.

T
he
result of

the

first

experiment

as same as his previous study
in 1997

[17]
.

T
he result of
second
experiments showed that there is no way to measure one algorithm is better
than another in
understanding point of view.


I
n summary,
the most important drawing aesthetic is minimizing edge crossings.
M
oreover, minimizing the bends is not suitable for all the
graphs

and maximizing
perceptual symmetries

is somewhat not effectiveness of improving the understanding
of graph.






-

14
-


3.

M
ethodology


3.1

Methods Adopted


This study

is
based

on

basic drawing skills, the original graph layout

algorithm
s

and
the most effective graph aesthetic

and the zoom and pan technique.

T
hus
,

all the basic
drawing skills will be adopted in this study

for drawing the graph

and
zoom and pan
technique

will be

ado
pted in this study

for graph navigation and interaction
.

B
ased on
the above discussion,
minimizing the edge crossings is
the most effectively graph
aesthetic

so minimizing the edge crossings will be
adopted

in this study
for
designing
the layout algorithm
and measuring the quality of the layout algorithm.

F
or the
original graph layout algorithms,
sugiyama layout algorithm will be modified and
adopted in this study.


M
y
research
question
is to design

the layout algorithm of relational database
representation

to draw a quality graph of relational database which can aid user
studying and using relational database more effectively


3.2 Research Process


T
his research

will be conducted in the following steps:



1
.
I will create a
relatively simple
test
relational database
at first
.


-

15
-


T
he purpose of this step

is obvious that
the simple test relational database let me
to test the layout algorithm on it.
I
f the layout algorithm cannot present the
quality graph, it also cannot work on the normal relational da
tabase.




2
.
I will initialize a
logical
graph

and this graph will
enlarge by itself if any

drawing on this graph outside the border of the graph.

T
he purpose of this step is to represent the drawing on the large enough graph and
then I can translate it
to a relatively small graph in the latter

step.




3
.

G
et the data information from the test database.




4
.

R
epresent the test database by using designed layout algorithm.

In this step, exactly, conclude
s

two steps:
first, design the algorithm; second, use
it

to represent the test database.


For designing the algorithm step, I propose
the first two steps of
two initial basic
ideas for designing algorithm.

1.

I
n each step choosing the node(s) which has(have) most number of
relations; if there are two or more nodes

have the same number, testing if
these nodes have relations with the nodes which have already been
selected
……

2.

In first step, choosing the
node(s) which has
(have) most number of
relations and positioning it on the left of the graph, calling it N1;

-

16
-


I
n secon
d step, choosing t
he node(s) which has(have) relation with N1 has
(have) most number of relations and positioning it far away NI

.




5
. Evaluate the designed algorithm and revise it if necessary



6
.

R
edo the 4
th

and 5
th

step until get the best algorithm.




7.

When I get the best algorithm, the algorithm plans to be implemented by Java.
T
hen, we can use this program to get the data from the relational database and
output a quality graph that represents the relational database.


3.3 Expected Outcome


I will prov
ide the very simple
test
relational database and show the expected outcome
for this database by using the first basic algorithm I discuss above
,
choosing

the most
number of relations nodes.


T
he simple test relational database

is as
follows
:

The
underlines

are used to denote the attributes of a PK
.

Dependent(
Employee
,
FirstName, Age, Sex,):
FK=(Employee)~>Employee(EmployeeNo)

Employee(
Empl
oy
eeNo
, LastName, FirstName, Sex, Age, PhoneNo, Salary,
Department):
FK=(Department)~>Department(DeptNo)

Division(
Name
,
City, State, Director,): FK=(Director)~>Employee(EmployeeNo)

-

17
-


Education(
Employee
,
School
,

Degree
, BeginYear, FinYear, GPA):
FK=(Employee)~>Employee(EmployeeNo)

Department(
DeptNo
, DeptName, City, State, Budget, Manger, Division):
FK1=(
Manger
)
~>Employee(Emplo
yeeNo) FK2=(Division)~>Division(Name)


The expected representation of this simple test database by using the first basic
algorithm is as follows

Graph1
:

Division
Dependent
Employee
Education
Department
FK
=
Employee
(
Department
)
~>
Department
(
DeptNo
)
FK
1
=
Department
(
Manger
)
~>
Employee
(
EmployeeNo
)

Graph 1

In the Graph 1, the five vertices mean the five
relational tables and the edges mean the
relations between the tables.
When we move the mouse to the edges, we can view the
relation details as aboved.




-

18
-










4.

Timetable

Study Period 5, 2010

Week

Date

Activities

Week 1

26
-
Jul
-
10



Develop the
test
database

dataset.

Week 2

02
-
Aug
-
10



Developing and performing the
algorithm

Week 3

09
-
Aug
-
10



R
evise the algorithm



Developing and performing the
revised
algorithm



Writing

thesis
-

introduction

Week 4

16
-
Aug
-
10



R
evise the algorithm



Developing and performing the
revised algorithm



W
riting thesis


relational database representation

Week 5

23
-
Aug
-
10



Get the best algorithm



W
riting thesis


literature review

Week 6

30
-
Aug
-
10



W
riting thesis


minimizing crossings algorithm

Week 7

06
-
Sep
-
10



W
riting thesis


methodology

Week 8

13
-
Sep
-
10



W
riting thesis


results and analysis

-

19
-


T
eaching
break

20
-
Sep
-
10



W
riting thesis


conclusion and future work

T
eaching
break

27
-
Sep
-
10



W
riting thesis


first draft to

superviso
r

Week 9

04
-
Oct
-
10



F
irst draft feedback



R
evise thesis

Week 10

11
-
Oct
-
10



S
econd draft to supervisor

Week 11

18
-
Oct
-
10



S
econd draft feedback



P
resentation slide to superviso
r

Week 12

25
-
Oct
-
10



The finalization of the thesis



P
resentation feedback



H
and final thesis

5.

R
eference

[1]

G. D. Battista
, et al.
, "Algorithms for drawing graphs: an annotated
bibliography,"
Computational Geometry,
vol. 4, pp. 235
-
282, 1994.

[2]

G. D. Battista
, et al.
,
Graph Drawing: Algorithms for the Visualization of
Graphs
: Prentice Hal
l PTR, 1998.

[3]

F. Bertault, "A force
-
directed algorithm that preserves edge crossing
properties," in
Graph Drawing
. vol. 1731/1999, ed Berlin / Heidelberg:
Springer, 1999, pp. 351
-
358.

[4]

P. Chen, "The entity
-
relationship model

toward a unified view of
data,"
ACM
Transactions on Database Systems (TODS),
vol. 1, p. 36, 1976.

[5]

E. Codd, "A relational model of data for large shared data banks,"
Communications of the ACM,
vol. 13, p. 387, 1970.

[6]

P. Eades, "A heuristic for graph drawing,"
Congressus nume
rantium,
vol. 42,
pp. 194
-
202, 1984.

[7]

P. Eades, "Drawing free trees,"
Bulletin of the Institute for Combinatorics and
its Applications,
vol. 5, pp. 10
-
36, 1992.

[8]

P. Eades and K. Sugiyama, "How to Draw a Directed Graph,"
Information
Processing,
vol. 1
3, pp. 424
-
434, 1990.

-

20
-


[9]

N. Gershon
, et al.
, "Information visualization,"
interactions,
vol. 5, pp. 9
-
15,
1998.

[10]

I. Herman
, et al.
, "Graph visualization and navigation in information
visualization: A survey,"
Visualization and Computer Graphics, IEEE
Transactions on,
vol. 6, pp. 24
-
43, 2000.

[11]

S. Juhn and J. Naumann, "The effectiveness of data representation
characteristics on user validation," 1985, pp. 212

226.

[12]

W. Kenny Jih
, et al.
, "The effects of relational and entity
-
relationship data
mode
ls on query performance of end users,"
International Journal of
Man
-
Machine Studies,
vol. 31, pp. 257
-
267, 1989.

[13]

R. Leitheiser and S. March, "The influence of database structure representation
on database system learning and use,"
Journal of Managemen
t Information
Systems,
vol. 12, p. 213, 1996.

[14]

Q. Li and Y. Chen, "Entity
-
Relationship Diagram,"
Modeling and Analysis of
Enterprise and Information Systems,
pp. 125
-
139, 2010.

[15]

F. Lochovsky and D. Tsichritzis, "User performance considerations in
DBMS
selection," 1977, p. 134.

[16]

P. Mutzel
, et al.
, "A library of algorithms for graph drawing," 1998, pp.
456
-
457.

[17]

H. Purchase, "Which aesthetic has the greatest effect on human
understanding?," in
Graph Drawing
. vol. 1353/1997, ed Berlin / Heidel
berg:
Springer, 1997, pp. 248
-
261.

[18]

H. Purchase
, et al.
, "Validating graph drawing aesthetics," in
Graph Drawing
.
vol. 1027/1996, ed Berlin / Heidelberg: Springer, 1995, pp. 435
-
446.

[19]

H. C. Purchase, "Effective information visualisation: a study of

graph drawing
aesthetics and algorithms,"
Interacting with Computers,
vol. 13, pp. 147
-
162,
2000.

[20]

H. C. Purchase, "Metrics for Graph Drawing Aesthetics,"
Journal of Visual
Languages & Computing,
vol. 13, pp. 501
-
516, 2002.

[21]

H. C. Purchase
, et al.
, "An experimental study of the basis for graph drawing
algorithms,"
J. Exp. Algorithmics,
vol. 2, p. 4, 1997.

[22]

E. Reingold and J. Tilford, "Tidier drawings of trees,"
IEEE Transactions on
Software Engineering,
vol. 7, pp. 223
-
228, 1981.

[23]

G. Robert
son
, et al.
, "Cone trees: animated 3D visualizations of hierarchical
information," 1991, p. 194.

-

21
-


[24]

M. Sharirk, "Quasi
-
Planar Graphs Have a Linear Number of Edges," 1995.

[25]

Y. Shiloach, "Arrangements of planar graphs on the planar lattices,"
Rehovot,
Israel: Weizmann Institute of Science,
1976.

[26]

K. Sugiyama
, et al.
, "Methods for visual understanding of hierarchical system
structures,"
IEEE TRANS. SYS. MAN, AND CYBER.,
vol. 11, pp. 109
-
125,
1981.