A Research Framework for Empirical Software Engineering Collaboration and Its Application in a Software Development Project

1
A Research Framework for Empirical Software
Engineering Collaboration and Its Application in
a Software Development Project
Software Engineering Center
Information-technology Promotion Agency, Japan

• Yoshiki Mitani, Nahomi Kikuchi, Tomoko
  Matsumura, Satoshi Iwamura,
• Mike Barker, Ken-ichi Matsumoto
• IPA/SEC,NAIST(EASE),IPA/SEC,NAIST(EASE),
  NTT Software(EASE)

Demonstrate and verify empirical software
engineering methods A new research framework
In response to government policy
Field data sharing between Industry Academia
Concept of a Macro-measurement
tools Target project A Governmental
Multi-Vendor Consortium type Software Project
IPA Information Technology Promotion Agency,
Japan SEC Software Engineering Center
EASE Empirical Approach to Software Engineering
project NAIST Nara Institute of Science and
Technology
2
In-process Project Measurement and Feedback
Structure
Feedback to in-process Software Development
Project
EPM Basic Analyzer
RDB
Development Environment
EPM Extended Analyzer
XML Standard Data Format
Configuration Management System
Program Development
Operation Rile
(CVS)
Code Clone Analyzer
Source Code
Bug Report
Bug Tracking System
(GNATS)
Bug Management
40 Management Items Operation Rule, Input Web
Template
Mailing-list Management System
Mail
e-Mail
Operation Rule
Review Report Sheet
Review
20 Management Items Operation Rule, Input
Template
Collaborating Filtering tool
400 Management Items
Benchmark Data Report
SEC Benchmark Database (over 1,000projects)
400 Items
QA Interview to PM
Interview Check-list Analyzer
QA Check-list
3
Table1 Measurement Data Items
Proceeding Project Data Data item Review record, Bug report(40 items), Specification Change report, Inter company mail with basic document, Basic Design Document, CVS repository with source code
Proceeding Project Data Tool CVS, GNATS, Mailman, EPM Data input form template Review report form, GNATS web template form
Proceeding Project Data Operation rule CVS, GNATS, Mail, EPM Collected data are kept safely in SEC secret data room.
Proceeding Project Data Data collecting Method Individual IT vender development process Weekly collect data by media from CVS and GNATS. Integration test process Use single GNATS system and collect data from it. CVS data are same as individual development process.
Ended Project Data/ Bench mark Data Data item SEC defined 400 bench mark data items, IT skill data based on IPA/ITSS
Ended Project Data/ Bench mark Data Data collecting Method SEC defined bench mark data input form( EXCEL template) Data collection timing end of basic design and end of development
Project Context Monitoring project Observe whole development process, Join to project meeting.
EPM Empirical Project Monitor ITSS Information
Technology Skill Standard
4
Table 2 Data Analysis Example
Basic Analysis Source line of code transition, Check-in opportunity, Check-out opportunity frequency, Number opportunity of mail, Number opportunity of bur report, Pending bug transition, Residual bug transition,
Extended Analysis Review quality analysis from IEEE 982.1(1988)Fault Density, Defect Density, Measurement from PSP (Personal Software Process) ODC (Orthogonal Defect Classification) Review quality File renewal analysis
Bench mark Analysis Source Code Clone Analysis Retrieve similar project from 1,000 past project benchmark data .
Skill data Analysis Analyze human factor from IT skill data of development persons
Code Clone Analysis
5
Fig.1 Macro Measurement Tools and its
Management Operation for Software Project
Measurement
ASDP Advanced Software Development Project
Measurement Target Project Software
Tools Measurement Platform Industry
Academia Collaboration Framework Ministry
Office Policy Government Policy
Development Project Management Operation
CVS, GNATS, Mailman
Software Tool Management Operation
EPM,400itemBenchmarkData,ITSS
Measurement Platform Management Operation
Macro Measurement Tools
EASE project, SEC
AI Collaboration Management Operation
MEXT, METI
National Policy
Trigger Budget and Policy
Prime Minister Leadership
EPM Empirical Project Monitor ITSS Information
Technology Skill Standard EASE Empirical
Approach to Software Engineering SEC Software
Engineering Center
MEXT Ministry of Education, Culture, Sports,
Science and Technology METI Ministry of
Economy, Trade and Industry
6
The major factors to realize measurement and
data-sharing

• The fundamental IA collaboration research
  framework
• Concept and equipment of measurement platform
• Thoughtful measurement plan
• Total framework coordination

Acknowledgment This work is supported by
IPA/SEC, METI and MEXT of Japan as part of the
Comprehensive Development of e-Society Foundation
Software program. We thank the researchers in
SEC and the EASE project who kindly support our
project.
Software Engineering Center
Information-technology Promotion Agency, Japan