How History Justifies System Architecture (or Not)

1/12

�

�

�

�

�

�

International Workshop on Principles of Software Evolution · Helsinki, Finland, 1 September 2003

How History JustifiesSystem Architecture (or not)

Thomas Zimmermann(with Stephan Diehl and Andreas Zeller)

Lehrstuhl SoftwaretechnikUniversitat des Saarlandes, Saarbrucken, Germany

2/12

�

�

�

�

�

�

The Problem

Your task: extend the debug component in GCC!

You identify the variable xcoff debug hooks.

What else do you need to change?

2/12

�

�

�

�

�

�

The Problem

Your task: extend the debug component in GCC!

You identify the variable xcoff debug hooks.

What else do you need to change?

General issue: only change coupled entities!

You can detect existing coupling by

• Program Analysis—e.g. def-use associations.

• Learning from History—entities changed together.

3/12

�

�

�

�

�

�

Evolutionary Coupling

gcc/gcc/dbxout.c [134] gcc/gcc/sdbout.c [74]34

dbx_debug_hooks

xcoff_debug_hooks

sdb_debug_hooks

[12]

[10]

[12]

3/12

�

�

�

�

�

�



dbx_debug_hooks

xcoff_debug_hooks

sdb_debug_hooks

[12]

[10]

[12]12

10 10

dbx_debug_hooks

xcoff_debug_hooks

sdb_debug_hooks

[12]

[10]

[12]

Support: How much evidence (= simultaneous changes)?Confidence: How relevant is coupling for participants?

3/12

�

�

�

�

�

�



dbx_debug_hooks

xcoff_debug_hooks

sdb_debug_hooks

[12]

[10]

[12]

4

4

xcoff_debug_hooks sdb_global_decl()

[4]

4

dbx_functions_end()

[7]

dbx_symbol_name()

[6] 2

12

10 10

dbx_debug_hooks

xcoff_debug_hooks

sdb_debug_hooks

[12]

[10]

[12]

Support: How much evidence (= simultaneous changes)?Confidence: How relevant is coupling for participants?

4/12

�

�

�

�

�

�

What We Do

Our ROSE prototype analyzes evolution of CVS archives.

ROSE

Step 1: Restore Transactions from CVS

Step 2: Identify Modified Entities

Reengineering Of Software EvolutionCVS

Metrics

Couplings

Graphs

ROSE determines entities at different granularities:

coarse-granular entities: directories, modules, files

fine-granular entities: methods, variables, sections

5/12

�

�

�

�

�

�

Step 1: Restoring Transactions

Two atomic changes δi and δi+1 are part of onetransaction ∆ = (δ1, . . . , δn) if:

author(δi) = author(δi+1)∧log message(δi) = log message(δi+1)∧

|time(δi+1)− time(δi)| < 〈200 seconds〉

We use a sliding window instead of a fixed one.

GNU C Compiler (GCC):The average transaction length is 6.2 seconds.The maximal transaction length is 1 hour 32 minutes.

6/12

�

�

�

�

�

�

Step 2: Light-Weight Analysis

File: Animals.java

1

3

23

25

30

56

58

60

80

99

public String[] COLORS = {


class Cat {

public Cat() {

...}

...}

...}

class Dog {

...}

...}

6/12

�

�

�

�

�

�


File: Animals.java

1

3

23

25

30

56

58

60

80

99

Cat.COLORSlines 3-23

Cat.Cat()lines 25-30

Class Catlines 1-56

Dog.COLORSlines 60-80

Class Doglines 58-99

Step A: Map to Entities



class Cat {

public Cat() {

...}

...}

...}

class Dog {

...}

...}

6/12

�

�

�

�

�

�


File: Animals.java

1

3

23

25

30

56

58

60

80

99



Class Catlines 1-56



Step A: Map to Entities



Class Catlines 1-56



Step B: Filter Entities

17public String[] COLORS = {


class Cat {

public Cat() {

...}

...}

...}

class Dog {

...}

...}

We analyze C/C++, JAVA, PYTHON, TEX and TEXINFO files.We get the modified methods, variables and subsections.

7/12

�

�

�

�

�

�

Example: GCC

i386_costk6_cost

i486_cost

pentium_cost

pentiumpro_cost

[11]

[11]

[12]

[12]

[14]

i386.c

7/12

�

�

�

�

�

�

Example: GCC

11

i386_costk6_cost

i486_cost

pentium_cost

pentiumpro_cost

[11]

[11]

[12]

[12]

[14]

i386.c

7/12

�

�

�

�

�

�

Example: GCC

11

i386_costk6_cost

i486_cost

pentium_cost

pentiumpro_cost

[11]

[11]

[12]

[12]

[14]

i386.c

processor_cost

[11]

i386.h

9

8/12

�

�

�

�

�

�

Visualizing Coupling

A DCB

A

D

C

B

High Confidence

Low ConfidenceNo Coupling (No Support)

8/12

�

�

�

�

�

�

Visualizing Coupling

A DCB

A

D

C

B

High Confidence

Low ConfidenceNo Coupling (No Support)

[10] [4][3]

A CA⇒ C: Confidence 3/10 = 30%C ⇒ A: Confidence 3/4 = 75%

9/12

�

�

�

�

�

�

Comparing Architecture with Evolution

DDD Source

LibrariesPicsIcons

Patches

Tests

Directoryddd/

9/12

�

�

�

�

�

�

Comparing Architecture with Evolution

Bad architecture

Better architecture

DDD Source

LibrariesPicsIcons

Patches

Tests

Directoryddd/

10/12

�

�

�

�

�

�

Measuring Evolutionary Coupling

Evolutionary Coupling Index (ECI).Different levels: entity/file or file/directory level.

ECI = # external couplings# internal couplings

The lower the ECI, the better the modularity.

10/12

�

�

�

�

�

�

Measuring Evolutionary Coupling

Evolutionary Coupling Index (ECI).Different levels: entity/file or file/directory level.

ECI = # external couplings# internal couplings

The lower the ECI, the better the modularity.

File/Directory Entity/FileECI ECI ECIfiltered

GCC 5.757 3.615 1.504DDD 0.250 4.462 1.922APACHE 2.827 11.815 0.675OPENSSL 8.665 101.053 7.859

Comparing only one level may be misleading (DDD).

11/12

�

�

�

�

�

�

Guiding the Programmer

12/12

�

�

�

�

�

�

Conclusion

Fine-grained evolutionary coupling. . .

• detects coupling between non-program entities.e.g. coupling between a function and a database schema

• guides developers while making changes.Programmers who changed this function also changed. . .

• gives better(?) results than coarse-grained coupling.

Coupling between files doesn’t tell you that much

• can be compared with given coupling (= architecture).

Results are mixed—what is coupling, anyway?

Those who cannot learn from history are doomed to repeat it.(George Santayana)

How History Justifies System Architecture (or Not)

Technology

public cat

debug component

public string colors

entities1 class cat

java1 class cat

variable xco debug hooks

class cat25lines

colorsclass dog