The Good the Bad and the Ugly of Dealing with Smelly Code (ITAKE Unconference)

The Good the Bad and the Ugly of
Dealing with Smelly Code
Dr. Radu Marinescu
radum@intooitus.com
IT.A.K.E - May 30, 2013

Full Professor. PhD Advisor
software engineering
since 2013

software maintenance. quality assurance
Co-founder. Head
since 2003

ECOOP 1998 WCRE 2005WCRE 2004 SYNASC 2005 ECOOP 1998 WCRE 2005WCRE 2004 SYNASC 2005

500+
citations
in mainstream
publications

Michele Lanza
Radu Marinescu
Object-Oriented
Metrics
in Practice
Using Software Metrics to
Characterize, Evaluate, and Improve
the Design of Object-Oriented Systems
Foreword by Stéphane Ducasse
Lanza·Marinescu
Object-Oriented
MetricsinPractice
tant Professor at the University of Lugano, Switzer-
rests lie in software (re)engineering and software evo-
n software visualization and metrics. He is the creator
ailable language-independent software visualization
e Ernst Denert Software Engineering Award in 2003.
istant Professor at the Politehnica University of
earch focus is on object-oriented design, reenginee-
He is the author of a suite of novel object-oriented
or of iPlasma, an integrated and freely available tool-
. Several of his published research ideas have already
own “Borland Together Control Center” CASE tool.
ented Metrics in
very engineering discipline. However, due to its lack
mplexity, software engineering is not considered a
y. Moreover, defining, understanding and applying
s like an overly complex activity, recommended only
general, if a software system is delivering the
few people – if any – care about measuring the qua-
Consequently, software metrics are still regarded
st software developers.
stify the design metrics used to assess the size,
bject-oriented software systems. Based on a novel
ally accepted semantics for metrics and by statistical
ustrial projects, they deduce a suite of metrics-based
esign of object-oriented software systems. They
fy design disharmonies in code, and how to devise
cally sound results and practically tested procedu-
es this book an ideal companion for professional
pers and quality engineers. The pattern-oriented des-
ers easy access to detecting shortcomings and
oblems.
ect-oriented
*many reengineering strategies
for poorly structured code
*brief introduction to software
visualization
‘’This well-written book is an impor-
tant piece of work that takes the
seemingly forgotten art of object-
oriented metrics to the next level in
terms of relevance and usefulness.’
Richard C. Gronback,
Chief Scientist, Borland Software
Corporation
1 3
1

Michele Lanza
Radu Marinescu
Object-Oriented
Metrics
in Practice
Using Software Metrics to
Characterize, Evaluate, and Improve
the Design of Object-Oriented Systems
Foreword by Stéphane Ducasse
Lanza·Marinescu
Object-Oriented
MetricsinPractice
tant Professor at the University of Lugano, Switzer-
rests lie in software (re)engineering and software evo-
n software visualization and metrics. He is the creator
ailable language-independent software visualization
e Ernst Denert Software Engineering Award in 2003.
istant Professor at the Politehnica University of
earch focus is on object-oriented design, reenginee-
He is the author of a suite of novel object-oriented
or of iPlasma, an integrated and freely available tool-
. Several of his published research ideas have already
own “Borland Together Control Center” CASE tool.
ented Metrics in
very engineering discipline. However, due to its lack
mplexity, software engineering is not considered a
y. Moreover, defining, understanding and applying
s like an overly complex activity, recommended only
general, if a software system is delivering the
few people – if any – care about measuring the qua-
Consequently, software metrics are still regarded
st software developers.
stify the design metrics used to assess the size,
bject-oriented software systems. Based on a novel
ally accepted semantics for metrics and by statistical
ustrial projects, they deduce a suite of metrics-based
esign of object-oriented software systems. They
fy design disharmonies in code, and how to devise
cally sound results and practically tested procedu-
es this book an ideal companion for professional
pers and quality engineers. The pattern-oriented des-
ers easy access to detecting shortcomings and
oblems.
ect-oriented
*many reengineering strategies
for poorly structured code
*brief introduction to software
visualization
‘’This well-written book is an impor-
tant piece of work that takes the
seemingly forgotten art of object-
oriented metrics to the next level in
terms of relevance and usefulness.’
Richard C. Gronback,
Chief Scientist, Borland Software
Corporation
1 3
1
1000+reprinted2010

Innovation Award 2006
2006
Continuous detection of design problems

“...have done the most to improve programmer productivity ...”
IBM
2009

“...have done the most to improve programmer productivity ...”
IBM
jury included
Grady Booch, Erich Gamma,
Fred Brooks, Frances Allen
2009

WorkingConferenceon
Reverse Engineering
InternationalConferenceon
Program Comprehension
ConferenceonSoftware
Maintenance & Re-Engineering

International Conference on Software Maintenance
2010
General Chair

T.J. Watson Research
Invited Talks

Dr. Radu MarinescuDr. Adrian Trifu Dr. Mircea Trifu George Ganea Ioana Verebi

Dr. Radu MarinescuDr. Adrian Trifu Dr. Mircea Trifu George Ganea Ioana Verebi
...uniquely innovative tools, and project-speciﬁc consultancy
to support complex quality assessment tasks
in large-scale software systems

inFusion
control the architecture and design quality of your system!
http://www.intooitus.com/products/infusion

inFusion
1+ Billion LOC total
10.000+ users

your daily companion for code and design quality
http://www.intooitus.com/products/incode
inCode

your daily companion for code and design quality
http://www.intooitus.com/products/incode
inCode
New version
launched yesterday!

Software is complex.
The Standish Group, 2004
53% Challenged
18% Failed
29% Succeeded

}
{
}
{
}
{
}
{
}
{
How complexis your project?

1’000’000 lines of code
* 3 = 3’000’000 seconds

* 3 = 3’000’000 seconds
/ 3600 = 833 hours

* 3 = 3’000’000 seconds
/ 3600 = 833 hours
/ 8 = 104 days

* 3 = 3’000’000 seconds
/ 3600 = 833 hours
/ 8 = 104 days
/ 22 = 4.5 months

But, code is for the computer.
Why would we ever read it?

forward
engineering
}
{
}
{
}
{
}
{

forward
engineering
actual development }
{
}
{
}
{
}
{
}
{
}
{
}
{
}
{
}
{

forward
engineering
actual development }
{
}
{
}
{
}
{
}
{
}
{
}
{
}
{
}
{
The only real documentation:
the code!

Ward Cunningham, 1992
technical debt.

When, due to constraints,
I design quickly and dirty,
my project gets loaded with
Ward Cunningham, 1992
technical debt.

You cannot control
what you cannot measure.
Tom de Marco

1Measure
design
2Assess
design
3Improve
design

Software metrics are measurements which
relate to software systems, processes or
related documents.

Metrics compress system traits into numbers.

Lorenz, Kidd, 1994
Chidamber, Kemerer, 1994

Lorenz, Kidd, 1994
LOC - number of lines of code
CYCLO - cyclomatic complexity of a function
NOF - number of functions
FANOUT - outgoing coupling

NOA - number of attributes
DIT - depth of inheritance tree
TCC - tight class cohesion
Lorenz, Kidd, 1994
LOC - number of lines of code
CYCLO - cyclomatic complexity of a function
NOF - number of functions
FANOUT - outgoing coupling

Metric Value
LOC (lines of code) 8.000
Would you maintain this system ?

Metric Value
NOF (functions) 2

Metric Value
NOF (functions) 2
Changed your mind, right? ;-)

Metric Value
LOC 35175
NOM 3618
NOC 384
CYCLO 5579
NOP 19
FANOUT 8590

Metric Value
LOC 35175
NOM 3618
NOC 384
CYCLO 5579
NOP 19
FANOUT 8590
?
Are these numbers “normal” ?

0.31ANDC
NOM
20.21 19
0.12
35175
NOP
NOC
418
0.15
8590
LOC
3618
9.42
5579
NOM
CALLS15128
384
FANOUT
9.72
0.56
AHH
CYCLO
Overview Pyramid provides a system’s overview.
Marinescu, Lanza

0.31ANDC
NOM
20.21 19
0.12
35175
NOP
NOC
418
0.15
8590
LOC
3618
9.42
5579
NOM
CALLS15128
384
FANOUT
9.72
0.56
AHH
CYCLO
Size
Marinescu, Lanza

0.31ANDC
NOM
20.21 19
0.12
35175
NOP
NOC
418
0.15
8590
LOC
3618
9.42
5579
NOM
CALLS15128
384
FANOUT
9.72
0.56
AHH
CYCLO
Size Communication
Marinescu, Lanza

0.31ANDC
NOM
20.21 19
0.12
35175
NOP
NOC
418
0.15
8590
LOC
3618
9.42
5579
NOM
CALLS15128
384
FANOUT
9.72
0.56
AHH
CYCLO
Size Communication
Inheritance
Marinescu, Lanza

0.31ANDC
NOM
20.21 19
0.12
35175
NOP
NOC
418
0.15
8590
LOC
3618
9.42
5579
NOM
CALLS15128
384
FANOUT
9.72
0.56
AHH
CYCLO
Marinescu, Lanza
More numbers?!?

...
HIGH
0.30
16
15
10
9
0.25
AVG
C++
4
5
0.20
LOW
Java
AVGLOW HIGH
0.24
10
13
7
0.20
10
0.16
7
4NOM/NOC
LOC/NOM
CYCLO/LOC

...
HIGH
0.30
16
15
10
9
0.25
AVG
C++
4
5
0.20
LOW
Java
AVGLOW HIGH
0.24
10
13
7
0.20
10
0.16
7
4NOM/NOC
LOC/NOM
CYCLO/LOC
4.000+OSS projects
500.000.000+LOC

...
HIGH
0.30
16
15
10
9
0.25
AVG
C++
4
5
0.20
LOW
Java
AVGLOW HIGH
0.24
10
13
7
0.20
10
0.16
7
4NOM/NOC
LOC/NOM
CYCLO/LOC
4.000+OSS projects
500.000.000+LOC
...and counting...

0.31ANDC
NOM
20.21 19
0.12
35175
NOP
NOC
418
0.15
8590
LOC
3618
9.42
5579
NOM
CALLS15128
384
FANOUT
9.72
0.56
AHH
CYCLO
close to high close to average close to low
Marinescu, Lanza

close to high close to average close to low
Marinescu, Lanza

We measure to control quality!

We measure to detect abnormalities!

Imagine changing just a small design fragment

33%of all the classes
would require changes
Imagine changing just a small design fragment

Breaking design principles, rules and best practices
deteriorates the code;
it leads to design problems.

Foutse et. al. - An exploratory study of the impact of anti-patterns on class change- and fault-proneness, 2012

Foutse et. al. - An exploratory study of the impact of anti-patterns on class change- and fault-proneness, 2012
Classes participating in design problems are
signiﬁcantly more likely to be subject to changes and
to be involved in fault-ﬁxing changes (bugs)
“
”

God
Class
Brain
Class
Feature
Envy
Data
Class
Brain
Method
Significant
Duplication
Intensive
Coupling
Dispersed
Coupling
Shotgun
Surgery
Tradition
Breaker
Refused
Parent
Bequest
uses
has
is
has
has
has (partial)
is partially
has
is
is
has
Futile
Hierarchy
uses
has
has
is
has (subclass)
Riel, 1996
Brown, 1998
Fowler, 1999
Marinescu, 2004
Design Problems (anti-patterns)

How to detect design problems?

Treat design problems as ﬁrst-class entities

God Classes tend to centralize the intelligence of the
system, to do everything and to use data from small
data-classes.
A.Riel, 1996

God Classes tend
to centralize the intelligence of the system,
to do everything and
to use data from small data-classes.

God Classes
centralize the intelligence of the system,
do everything and
use data from small data-classes.

God Classes
are complex,
are not cohesive,
access external data.

God Classes
are complex,
are not cohesive,
access external data.
Compose metrics into queries using
logical operators

Detection Strategies are metric-based queries to
detect design ﬂaws.
METRIC 1 > Threshold 1
Rule 1
METRIC 2 < Threshold 2
Rule 2
AND Quality problem
Marinescu

A God Class centralizes too much intelligence in
the system.
ATFD > FEW
Class uses directly more than a
few attributes of other classes
WMC ≥ VERY HIGH
Functional complexity of the
class is very high
TCC < ONE THIRD
Class cohesion is low
AND GodClass
Marinescu

Envious Methods are more interested in data from
a handful of classes.
ATFD > FEW
Method uses directly more than
a few attributes of other classes
LAA < ONE THIRD
Method uses far more attributes
of other classes than its own
FDP ≤ FEW
The used "foreign" attributes
belong to very few other classes
AND Feature Envy
Marinescu, Lanza

Data Classes are dumb data holders.
WOC < ONE THIRD
Interface of class reveals data
rather than offering services
AND Data Class
Class reveals many attributes and is
not complex
Lanza, Marinescu 2006

Data Classes are dumb data holders.
AND
OR
Class reveals many
attributes and is not
complex
NOAP + NOAM > FEW
More than a few public
data
WMC < HIGH
Complexity of class is not
high
NOAP + NOAM > MANY
Class has many public
data
WMC < VERY HIGH
Complexity of class is not
very high
AND
Lanza, Marinescu 2006

20+ Detection strategies
for design ﬂaws on class, method and subsystem level

?Does anyone care design ﬂaws

63releases of Eclipse JDT and EMF

2
previous
release
9
current
release
7 3
3
6
5
Decaying
Class
Refactored
Class
NewFlawed
Class

0
2
4
6
8
10
1.1 2 2.1 2.2 2.3 2.4 2.5 2.6 2.7
0
2
4
6
8
10
1.1 2 2.1 2.2 2.3 2.4 2.5 2.6 2.7
0
2
4
6
8
10
1.1 2 2.1 2.2 2.3 2.4 2.5 2.6 2.7
0
2
4
6
8
10
1.1 2 2.1 2.2 2.3 2.4 2.5 2.6 2.7
0
2
4
6
8
10
1.1 2 2.1 2.2 2.3 2.4 2.5 2.6 2.7
0
2
4
6
8
10
1.1 2 2.1 2.2 2.3 2.4 2.5 2.6 2.7
0
2
4
6
8
10
2 2.1 3 3.1 3.2 3.3 3.4 3.5 3.6 3.7
0
2
4
6
8
10
2 2.1 3 3.1 3.2 3.3 3.4 3.5 3.6 3.7
0
2
4
6
8
10
2 2.1 3 3.1 3.2 3.3 3.4 3.5 3.6 3.7
0
2
4
6
8
10
2 2.1 3 3.1 3.2 3.3 3.4 3.5 3.6 3.7
0
2
4
6
8
10
2 2.1 3 3.1 3.2 3.3 3.4 3.5 3.6 3.7
0
2
4
6
8
10
2 2.1 3 3.1 3.2 3.3 3.4 3.5 3.6 3.7
JDT EMF
God
Class
Schizophrenic
Class
Refused
Parent
Bequest
Code
Duplication
Blob
Operation
Data
Clumps

0
2
4
6
8
10
1.1 2 2.1 2.2 2.3 2.4 2.5 2.6 2.7
0
2
4
6
8
10
1.1 2 2.1 2.2 2.3 2.4 2.5 2.6 2.7
0
2
4
6
8
10
1.1 2 2.1 2.2 2.3 2.4 2.5 2.6 2.7
0
2
4
6
8
10
1.1 2 2.1 2.2 2.3 2.4 2.5 2.6 2.7
0
2
4
6
8
10
1.1 2 2.1 2.2 2.3 2.4 2.5 2.6 2.7
0
2
4
6
8
10
1.1 2 2.1 2.2 2.3 2.4 2.5 2.6 2.7
0
2
4
6
8
10
2 2.1 3 3.1 3.2 3.3 3.4 3.5 3.6 3.7
0
2
4
6
8
10
2 2.1 3 3.1 3.2 3.3 3.4 3.5 3.6 3.7
0
2
4
6
8
10
2 2.1 3 3.1 3.2 3.3 3.4 3.5 3.6 3.7
0
2
4
6
8
10
2 2.1 3 3.1 3.2 3.3 3.4 3.5 3.6 3.7
0
2
4
6
8
10
2 2.1 3 3.1 3.2 3.3 3.4 3.5 3.6 3.7
0
2
4
6
8
10
2 2.1 3 3.1 3.2 3.3 3.4 3.5 3.6 3.7
JDT EMF
God
Class
Schizophrenic
Class
Refused
Parent
Bequest
Code
Duplication
Blob
Operation
Data
Clumps different
patterns

0
2
4
6
8
10
1.1 2 2.1 2.2 2.3 2.4 2.5 2.6 2.7
0
2
4
6
8
10
1.1 2 2.1 2.2 2.3 2.4 2.5 2.6 2.7
0
2
4
6
8
10
1.1 2 2.1 2.2 2.3 2.4 2.5 2.6 2.7
0
2
4
6
8
10
1.1 2 2.1 2.2 2.3 2.4 2.5 2.6 2.7
0
2
4
6
8
10
1.1 2 2.1 2.2 2.3 2.4 2.5 2.6 2.7
0
2
4
6
8
10
1.1 2 2.1 2.2 2.3 2.4 2.5 2.6 2.7
0
2
4
6
8
10
2 2.1 3 3.1 3.2 3.3 3.4 3.5 3.6 3.7
0
2
4
6
8
10
2 2.1 3 3.1 3.2 3.3 3.4 3.5 3.6 3.7
0
2
4
6
8
10
2 2.1 3 3.1 3.2 3.3 3.4 3.5 3.6 3.7
0
2
4
6
8
10
2 2.1 3 3.1 3.2 3.3 3.4 3.5 3.6 3.7
0
2
4
6
8
10
2 2.1 3 3.1 3.2 3.3 3.4 3.5 3.6 3.7
0
2
4
6
8
10
2 2.1 3 3.1 3.2 3.3 3.4 3.5 3.6 3.7
JDT EMF
God
Class
Schizophrenic
Class
Refused
Parent
Bequest
Code
Duplication
Blob
Operation
Data
Clumps
only
ﬂawed
by birth

0
2
4
6
8
10
1.1 2 2.1 2.2 2.3 2.4 2.5 2.6 2.7
0
2
4
6
8
10
1.1 2 2.1 2.2 2.3 2.4 2.5 2.6 2.7
0
2
4
6
8
10
1.1 2 2.1 2.2 2.3 2.4 2.5 2.6 2.7
0
2
4
6
8
10
1.1 2 2.1 2.2 2.3 2.4 2.5 2.6 2.7
0
2
4
6
8
10
1.1 2 2.1 2.2 2.3 2.4 2.5 2.6 2.7
0
2
4
6
8
10
1.1 2 2.1 2.2 2.3 2.4 2.5 2.6 2.7
0
2
4
6
8
10
2 2.1 3 3.1 3.2 3.3 3.4 3.5 3.6 3.7
0
2
4
6
8
10
2 2.1 3 3.1 3.2 3.3 3.4 3.5 3.6 3.7
0
2
4
6
8
10
2 2.1 3 3.1 3.2 3.3 3.4 3.5 3.6 3.7
0
2
4
6
8
10
2 2.1 3 3.1 3.2 3.3 3.4 3.5 3.6 3.7
0
2
4
6
8
10
2 2.1 3 3.1 3.2 3.3 3.4 3.5 3.6 3.7
0
2
4
6
8
10
2 2.1 3 3.1 3.2 3.3 3.4 3.5 3.6 3.7
JDT EMF
God
Class
Schizophrenic
Class
Refused
Parent
Bequest
Code
Duplication
Blob
Operation
Data
Clumps
strong
reaction
to debt

0
2
4
6
8
10
1.1 2 2.1 2.2 2.3 2.4 2.5 2.6 2.7
0
2
4
6
8
10
1.1 2 2.1 2.2 2.3 2.4 2.5 2.6 2.7
0
2
4
6
8
10
1.1 2 2.1 2.2 2.3 2.4 2.5 2.6 2.7
0
2
4
6
8
10
1.1 2 2.1 2.2 2.3 2.4 2.5 2.6 2.7
0
2
4
6
8
10
1.1 2 2.1 2.2 2.3 2.4 2.5 2.6 2.7
0
2
4
6
8
10
1.1 2 2.1 2.2 2.3 2.4 2.5 2.6 2.7
0
2
4
6
8
10
2 2.1 3 3.1 3.2 3.3 3.4 3.5 3.6 3.7
0
2
4
6
8
10
2 2.1 3 3.1 3.2 3.3 3.4 3.5 3.6 3.7
0
2
4
6
8
10
2 2.1 3 3.1 3.2 3.3 3.4 3.5 3.6 3.7
0
2
4
6
8
10
2 2.1 3 3.1 3.2 3.3 3.4 3.5 3.6 3.7
0
2
4
6
8
10
2 2.1 3 3.1 3.2 3.3 3.4 3.5 3.6 3.7
0
2
4
6
8
10
2 2.1 3 3.1 3.2 3.3 3.4 3.5 3.6 3.7
JDT EMF
God
Class
Schizophrenic
Class
Refused
Parent
Bequest
Code
Duplication
Blob
Operation
Data
Clumps
only
pay-back

Let’s start with a simple example...

public class TarHeader{
/**
* The entry's name.
*/
public StringBuffer name;
/**
* The entry's permission mode.
*/
public int mode;
/**
* The entry's user id.
*/
public int userId;
/**
* The entry's group id.
*/
public int groupId;
}

/**
* The entry's name.
*/
StringBuffer name;
/**
*/
int mode;
/**
*/
int userId;
/**
*/
int groupId;
}
public
public
public
public

/**
* The entry's name.
*/
StringBuffer name;
/**
*/
int mode;
/**
*/
int userId;
/**
*/
int groupId;
}
public
public
public
public
DATA
CLASS

/**
* The entry's name.
*/
StringBuffer name;
/**
*/
int mode;
/**
*/
int userId;
/**
*/
int groupId;
}

/**
* The entry's name.
*/
StringBuffer name;
/**
*/
int mode;
/**
*/
int userId;
/**
*/
int groupId;
}
private
private
private
private

/**
* The entry's name.
*/
StringBuffer name;
/**
*/
int mode;
/**
*/
int userId;
/**
*/
int groupId;
}
private
private
private
private
Encapsulate public data
(in TarHeader)1

Suddenly we have compile errors!

public class {
public void parseTarHeader( TarHeader hdr, byte[] header ){
int offset = 0;
hdr.name = TarHeader.parseName( header, offset,
TarHeader.NAMELEN );
offset += TarHeader.NAMELEN;
hdr.mode = (int)TarHeader.parseOctal( header, offset,
TarHeader.MODELEN );
offset += TarHeader.MODELEN;
hdr.userId = (int)TarHeader.parseOctal( header, offset,
TarHeader.UIDLEN );
offset += TarHeader.UIDLEN;
hdr.groupId = (int)TarHeader.parseOctal( header, offset,
TarHeader.GIDLEN );
}
TarEntry

public class {
int offset = 0;
TarHeader.UIDLEN );
TarHeader.GIDLEN );
}
TarEntry
TarHeader hdr
hdr.name
hdr.mode
hdr.userId
hdr.groupId
parseTarHeader

public class {
int offset = 0;
TarHeader.UIDLEN );
TarHeader.GIDLEN );
}
TarEntry
TarHeader hdr
hdr.name
hdr.mode
hdr.userId
hdr.groupId
parseTarHeader
Move the method
(TarEntry > TarHeader)1

public class {
int offset = 0;
TarHeader.UIDLEN );
TarHeader.GIDLEN );
}
TarEntry
TarHeader hdr
hdr.name
hdr.mode
hdr.userId
hdr.groupId
parseTarHeader
(in TarHeader)2
Move the method
(TarEntry > TarHeader)1

public class TarEntry{
int offset = 0;
hdr.name = TarHeader.parseName(
header, offset, TarHeader.NAMELEN );

hdr.mode = (int)TarHeader.parseOctal( header,
offset, TarHeader.MODELEN );
hdr.userId = (int)TarHeader.parseOctal( header,
offset, TarHeader.UIDLEN );
hdr.groupId = (int)TarHeader.parseOctal( header,
offset, TarHeader.GIDLEN );
}
public void writeEntryHeader(byte[] outbuf){
int offset = 0;
long size = this.header.size;
offset = TarHeader.getLongOctalBytes
( size, outbuf, offset, TarHeader.SIZELEN );
}
parseTarHeader
writeEntryHeader

int offset = 0;

}
int offset = 0;
}
parseTarHeader
writeEntryHeader
Duplicated
Code

int offset = 0;

}
int offset = 0;
}
parseTarHeader
writeEntryHeader
Extract method
(by factoring out the duplicated code)1

int offset = 0;

}
int offset = 0;
}
parseTarHeader
writeEntryHeader
Extract method
Move the newly created method
(in TarHeader)2

int offset = 0;

}
int offset = 0;
}
parseTarHeader
writeEntryHeader
Extract method
(in TarHeader)2
(in TarHeader)3

there is also an alternative solution ...

Move data
(TarHeader > TarEntry)1

Move data
(in TarEntry)2

vs.
Extract method
(out of the duplicated code)1
(in TarHeader)2
(in TarHeader)3
Move data
(in TarEntry)2

vs.
Which solution is best?
Extract method
(out of the duplicated code)1
(in TarHeader)2
(in TarHeader)3
Move data
(in TarEntry)2

Correction implies understanding the whole context
Contextual
1

Correction involves the decision of the engineer
Decisional
2

We need multiple instruments
to explore...

There are many ways to explore...

1854,
London,
cholera
epidemic

We are
visual
beings.
..and we’re good at
spotting patterns

Gestalt principles
proximity
enclosure connectivity
similarity

In 7 seconds count numbers larger than 0.85!
0.103 0.176 0.387 0.300 0.829 0.276 0.179 0.321 0.192 0.250
0.333 0.384 0.864 0.587 0.857 0.698 0.640 0.621 0.984 0.316
0.421 0.309 0.654 0.729 0.228 0.529 0.832 0.435 0.699 0.426
1.266 0.750 0.056 0.936 0.711 0.749 0.723 0.201 0.542 0.819
0.225 0.926 0.643 0.337 0.721 0.337 0.682 0.987 0.232 0.449
0.187 0.586 0.529 0.340 0.276 0.835 0.473 0.445 1.103 0.720
1.153 0.485 0.560 0.428 0.628 0.335 0.456 0.879 0.699 0.424

What about now?
0.103 0.176 0.387 0.300 0.829 0.276 0.179 0.321 0.192 0.250
0.333 0.384 0.864 0.587 0.857 0.698 0.640 0.621 0.984 0.316
0.421 0.309 0.654 0.729 0.228 0.529 0.832 0.435 0.699 0.426
1.266 0.750 0.056 0.936 0.711 0.749 0.723 0.201 0.542 0.819
0.225 0.926 0.643 0.337 0.721 0.337 0.682 0.987 0.232 0.449
0.187 0.586 0.529 0.340 0.276 0.835 0.473 0.445 1.103 0.720
1.153 0.485 0.560 0.428 0.628 0.335 0.456 0.879 0.699 0.424

What about now?
0.103 0.176 0.387 0.300 0.829 0.276 0.179 0.321 0.192 0.250
0.333 0.384 0.864 0.587 0.857 0.698 0.640 0.621 0.984 0.316
0.421 0.309 0.654 0.729 0.228 0.529 0.832 0.435 0.699 0.426
1.266 0.750 0.056 0.936 0.711 0.749 0.723 0.201 0.542 0.819
0.225 0.926 0.643 0.337 0.721 0.337 0.682 0.987 0.232 0.449
0.187 0.586 0.529 0.340 0.276 0.835 0.473 0.445 1.103 0.720
1.153 0.485 0.560 0.428 0.628 0.335 0.456 0.879 0.699 0.424
Pre-attentive features!

Remember...
* 3 = 3’000’000 seconds
/ 3600 = 833 hours
/ 8 = 104 days
/ 22 = 4.5 months

A picture is worth
a thousand words.
Proverb

Software visualization is more than UML!

PolymetricViews show up to 5 metrics.
Color
metric
Width metric
Height metric
Position metrics
Lanza, 2003

Example 1: System ComplexityView

System Complexity shows class hierarchies.
lines
attributes
methods
Lanza, Ducasse, 2003

Compresses 1.300 classes in one picture.

Code City shows where your code lives.
Wettel, Lanza, 2007
classes are buildings grouped in quarters of packages

Look carefully how the system evolves in time!

Polymetric Maps
display various design concerns, but reuse the layout!

Polymetric Maps
display various design concerns, but reuse the layout!
...and no edges!

Package Map: the layout
Eclipse-JDT, 2010

Package Map: design problems
Eclipse-JDT, 2010
severeno
affected by design problems

Package Map: coupling perspective
Eclipse-JDT, 2010
client (uses other classes)
provider (is used by other classes)
many
many
none
none

Eclipse-JDT, 2010

Eclipse-JDT, 2010
Interactive!

Multiple Inheritance ?
Interfaces?

severeno
affected by design problems

client (uses other classes)
provider (is used by other classes)
many
many
none
none

11.000+ classes in one picture...
Java JDK 1.5, 2011

A demo is worth
a thousand pictures.
Radu

inCode
intooitus.com/products/incode

The Good the Bad and the Ugly of Dealing with Smelly Code (ITAKE Unconference)

More Related Content

What's hot

Viewers also liked

Similar to The Good the Bad and the Ugly of Dealing with Smelly Code (ITAKE Unconference)

Recently uploaded

The Good the Bad and the Ugly of Dealing with Smelly Code (ITAKE Unconference)