Software modularity and reuse Paulo Borba Informatics Center Federal University of Pernambuco

[email protected] ◈ twitter.com/pauloborba Wednesday, August 18, 2010

1

Reuse opportunities

Wednesday, August 18, 2010

2

Different devices, 15 to 60 different applications…

Different clients, different products

http://www.androidauthority.com Wednesday, August 18, 2010

3

Little reuse and agility, high costs Even with J2ME!

Wednesday, August 18, 2010

4

Problem might also appear in the context of a single product... Wednesday, August 18, 2010

5

How to detect it? Wednesday, August 18, 2010

6

Code cloning tool, file

Wednesday, August 18, 2010

7

Code cloning tool, packages

Wednesday, August 18, 2010

8

But what is a clone? Wednesday, August 18, 2010

9

Clones? Strings? Tokens? if (first == null && second != null) { return -1; } if (first == null && second != null) { return -1; } if (f == null && s != null) { return -1; } if (f == null && s != null) { return 0; } Wednesday, August 18, 2010

10

Clones? Reuse opportunities? int getDay() { return day; } int getMonth() { return month; } int getDay() { return day+1; } Wednesday, August 18, 2010

11

Google wave clones int getDay() { return day; } int getMonth() { return month; } int getDay() { return day+1; } Wednesday, August 18, 2010

12

int sum(int x,int y){return x+y;} int sumInt(int a,int b){return a+b;} Wednesday, August 18, 2010

13

Changing clone notion, small

Wednesday, August 18, 2010

14

Changing clone notion, large

Wednesday, August 18, 2010

15

Minimum clone length (tokens) if (o.toString().equals("/")) { div = new DivExp(null, null)); } if (o.toString().equals("*")) { mult = new MultExp(null, null)); }

Wednesday, August 18, 2010

16

P-match... maps different variables and function names to different tokens if (x>0) {y=1; x=2; x=w;} else {y=1; x=2; x=z;} Wednesday, August 18, 2010

17

Minimum TKS (kinds of tokens) int x = 0; int y = 0; int z = 0; int a = 0; int b = 0; int c = 0;

Wednesday, August 18, 2010

18

Shaper level, hard and soft

Wednesday, August 18, 2010

19

Shaper level, easy and none

Wednesday, August 18, 2010

20

File metrics

 

• • •

LEN: File length (in tokens)

• • •

RSA: Ratio of similarity to another file

•

RNR: Ratio of non-repeated code

Wednesday, August 18, 2010

CLN: Number of Code Clones NBR: Neighbors, other files that share a code clone in this file

RSI: Ratio of similarity within the file CVR: Coverage, percentage of tokens covered by another code clone

21

Clone set metrics • • •

LEN: Length of a code fragment of the code clone

•

RAD (radius): Range of the source code fragments of a code clone in the directory hierarchy

•

RNR (ratio of non-repeated tokens): Ratio (percentage) of tokens that are not included in repeated part of a code fragment of the code clone

•

TKS (token set size): Size of a set of tokens of a code fragment of the code clone

•

LOOP, COND, McCabe

Wednesday, August 18, 2010

POP (population): Count of code fragments of the code clone NIF: Count of source files that include one or more code fragments of the code clone

22

Strategy • Start with stronger constraints • Analyze clones • Reduce constraints • Cycle until no more interesting clones Wednesday, August 18, 2010

23

Take notes, now! Wednesday, August 18, 2010

24

Reuse opportunities

Wednesday, August 18, 2010

25

Software modularity and reuse Paulo Borba Informatics Center Federal University of Pernambuco

[email protected] ◈ twitter.com/pauloborba Wednesday, August 18, 2010

26