Osaka University Applicant to ISERN

1
Osaka UniversityApplicant to ISERN

• Katsuro Inoue (Osaka Univ.)

2
Osaka University

• 70 years history
• 12 schools
• 20,000 students
• SE research lab 3 groups 9 faculty
  members

Osaka-u
NAIST
3
EASE Project Empirical Approach to Software
Engineering
4
What is EASE Project?

• Japans first large project for empirical
  software engineering (20032007 5 years)
• Government fund industry fund
• Leader Koji Torii, NAIST
• with many researchers in NAIST and Osaka
  Univ
• Laboratory at northern Osaka (Senri Lab)
• 1.5 hours from NAIST
• 15 minutes from Osaka Univ.

5
Goals

• Develop technologies and tools
• Effectively improve software productivity and
  quality
• Activate software industry
• Publish research papers
• Make bridge between software industry and
  academia

6
Classification of SE Technologies by Target Scale
7
Mega Software Engineering (MSE)

• Huge data collection from tens of thousands of
  projects
• Intensive analysis of the collected data
• Effective feedback for organizational benefit

8
Example Scenario of MSE (1)
Scheduled progress of project X
1
Actual progress of project X
2
Find projects similar to X - Project
categorization - Collaborative filtering
E
C
A
W
X
Y
V
Q
T
P
9
Example Scenario of MSE (2)
3
Average reuse rate in similar projects
Project Xs reuse rate
- Code-clone detection
Promote using software asset search engine to
project X
4
- Software asset search engine
10
MSE Framework
Analysis

Open-source Software


Data Collection
Feedback
Associate Organizations
Software Development Organization
11
Implementation of MSE Framework

• Open source development tools
• CVS versioning
• MailMan mail management
• Gnats issue tracking
• Existing Web interface
• Corporate Source (ZeeSource)

12
Standardizing MSE Data

• Product data
• CVS format
• Product definition in XML
• Process data
• Event sequence in XML
• XML DB
• PostgreSQL XML extension

13
MSE Technologies for Intensive Analysis(1)System
Classification by Similarity Measure
14
MSE Technologies for Intensive Analysis(2)
Finding the Most Similar Project by
Collaborative Filtering
Represen-tative
Focused
Collaborative
Q MResources
OutcomeAdopted
7.5 (target)
9
9
9
7
App. A
8
7
8
? (missing)
8
App. B
? (missing)
8
8
8
7
App. C
7
6
? (missing)
9
6
App. D
15
MSE Technologies for Intensive Analysis(3) Fast
Code Clone Detection by CCFinder
16
MSE Technologies for Intensive Analysis(4)
Software Asset Search by SPARS
17
(No Transcript)
18
(No Transcript)
19
Project Alliance
Technical Committees -Metrics -Security -Process
Improvement -Formal Method Associate
Companies Associate Universities
International Advisories Core Company -NTT
Software -Hitachi -Hitachi GP -SRA Core
Univ. -NAIST -Osaka Univ.
Senri Lab -Manager -Company engineers -Univ.
researchers
20
Future Plan

• Demonstrate the first prototype of MSE framework
  at the workshop in Nov. 2003
• Distribute the prototype to companies in 2004
• Organize a consortium for promotion

21

• END

22
(No Transcript)
23
Schedule

• First 3 years
• Make prototypes
• Evaluate the prototypes
• Complete distribution system
• Following 2 years
• Disseminate the system

24
System Architecture
Manager Developer
Analysis Tools(Next release)
Estimation Tools Similarity Tools ...
GUI
(5)
PostgreSQLXML Plug-in
(6) (6-2) (6-3) (6-4)
(5-0)
XML-DB Repository
Standardized Data
RubyScript
Translator1

Translator3
Translator2
GUI
Other Tools (Next release)
(2)
(3)
(4)
Option1
Option2
(1)
Manager Developer
CVS
Mail
Gnats(bug-rep.)
Log objects
CorporateSource
Software Development Environments
25
Java classes
Raw component archive
Component rank engine
Internet / Organization Repositories
Ranked component archive
Query handler
Query/ reply
SPARS-J
Software Component Searcher