Project Metrics - PowerPoint PPT Presentation

1 / 17
About This Presentation
Title:

Project Metrics

Description:

need to be careful that metrics measure concept of interest ... Function oriented metrics ... SEI Guidelines for Establishing a Metrics Program con't ... – PowerPoint PPT presentation

Number of Views:130
Avg rating:3.0/5.0
Slides: 18
Provided by: Gay2
Category:
Tags: metrics | project

less

Transcript and Presenter's Notes

Title: Project Metrics


1
Project Metrics
2
Overview
  • Motivation
  • Characterize
  • evaluate
  • predict
  • improve
  • software engineering
  • more difficult to quantify
  • lack of consensus

3
Overview cont
  • Goal - indicator
  • metric that provides insight
  • allow early detection of potential problems
  • need to be careful that metrics measure concept
    of interest
  • to measure process - must use data related to
    process

4
Overview cont
  • Public vs. Private use of metrics
  • private - only available to individual or group
  • public - usually summary of private information
  • metric (like any evaluation tool) - can be used
    to help promote improvement or input in personnel
    actions

5
Overview cont
  • Statistical software process improvement - uses
    software failure analysis to collect information
    about all errors defects
  • failure analysis
  • errors defects - categorized by origin
  • cost to correct each error defect - recorded
  • compare number of errors defects by category

6
Overview cont
  • compare cost of errors defects by category
  • analyze results to uncover problem categories
  • develop plans to modify process to reduce most
    costly errors
  • project metrics
  • used to modify workflow technical activities
  • give ongoing assessment of product quality

7
Software Measurements
  • Direct measures
  • measure quantity immediately
  • speed, size, ...
  • Indirect measures
  • ilities - functionality, reliability,
    maintainability, ...
  • Size-oriented metrics
  • need to normalize wrt size of project
  • common (but not best) approach - normalize wrt
    lines of code

8
Software Measurements cont
  • Function oriented metrics
  • measure of functionality - used as normalization
    value calculated from other measures
  • function point - related to achievement of
    specification
  • domain characteristics considered
  • number of user inputs
  • number of user outputs

9
Software Measurements cont
  • number of user inquiries
  • number of files
  • number of external interfaces
  • Complexity value - associated with each count
  • many methods of averaging these values
  • extended function point metrics
  • function point - inadequate for engineering
    embedded systems
  • feature point
  • takes into account complexity of algorithms

10
Software Measurements cont
  • relationship between lines of code function
    points - programming language dependent

11
Metrics for Software Quality
  • Need technical measures - not subjective measures
  • project scope - measure errors defects
  • errors per function point
  • errors uncovered per peer review
  • errors uncovered per testing hour

12
Metrics for Software Quality cont
  • Points of view of quality
  • product operation (using)
  • product modification (modifying)
  • product transition (porting)
  • indicators of quality
  • correctness
  • performs required functions
  • most common measure - KLOC

13
Metrics for Software Quality cont
  • Maintainability
  • ease with which programs can be modified
  • indirect measures used
  • MTTC (mean time to change)
  • spoilage - cost to modify
  • integrity
  • ability to withstand attacks to security
  • use attributes
  • threat - probability that an attack will occur
    within a specified time interval
  • security - probability that an attack will be
    repelled

14
Metrics for Software Quality cont
  • Integrity
  • usability
  • characteristics
  • physical intellectual skill required to learn
    system
  • time required to become moderately efficient
  • net increase in productivity for moderately
    efficient user
  • users attitudes
  • defect removal efficiency (DRE)
  • E - number of errors
    before delivery
  • D - number of
    errors after delivery

15
Reasons for Using Metrics
  • Without measurements - cannot tell whether
    improving
  • meaningful goals can be established

16
SEI Guidelines for Establishing a Metrics Program
  • Identify business goals
  • identify what want to know
  • identify subgoals
  • identify entities attributes related to
    subgoals
  • formalize measurement goals

17
SEI Guidelines for Establishing a Metrics Program
cont
  • identify quantifiable questions related
    indicators
  • Identify data elements to be collected
  • define measures to be used
  • identify actions to implement measures
  • prepare plan for implementing measures
Write a Comment
User Comments (0)
About PowerShow.com