Software Project Management SPM

1
Software Project Management (SPM)

• Lecture 4
• - Software Effort Estimate
• - Risk Assessments
• - Cost Benefit Analysis
• Dr. Daniel Keret

2
Reading Assignment

• Software Project Management, Bob Hughes and Mike
  Cotterell, McGraw-Hill, 3rd Edition.
• Chapters 2.8,3,5
• A Guide to the Project Management Body of
  Knowledge, PMI Publications, 3rd Edition, 2004
• Chapter 11.1

3
Software Effort Estimate

• Software Effort Estimation Techniques
• Function Points, Objects Points
• COnstructive COst MOdel (COCOMO)
• Expert Judgment
• On-Going Adjustments
• Strategic Planning, Feasibility Study, System
  Specification
• Project Planning, Programming Phases, Change
  Control
• Risk Assessments.

4
Accurate Estimation is a Management Goal

• Difficulties in adjusting previous projects
  actual effort.
• Major Projects are unique in nature
• Changing Technology
• Lack of homogenous previous projects data
• Subjective nature of estimates ( overestimate
  difficulties of small tasks, underestimate of
  large tasks as the complexity is not linear to
  project size)
• The need for accurate estimation
• Parkinsons Law Work expands to fill the time
  allocated.
• Brooks Law The work effort will grow
  disproportionately with the unneeded manpower
  assigned to the project.
• Weinbergs law of zero reliability If a system
  does not have to be reliable it can meet any
  other objective.
• Basic data for software estimating
• Historic Data
• Measure of Work (SLOC Source line s of code,
  Object, Function Points)
• Complexity

5
Estimation Techniques

• Bottom Up Adding the effort of each sub
  activity. (Work Breakdown Structure is needed.
  Good for later stages of the project)
• Top Down Needs Parametric Models (effort
  system size X productivity rate SLOC X lines of
  code per day per programmer)
• Expert Judgment Good for a Change in existing
  system.
• ( knowledge in the application area and the
  development environment)

6
Estimation Techniques (Cont.)

• Function Points FP (Evaluation of SLOC) Good
  for Information Systems.
• Counts
• external input types,
• external output types,
• logical internal record types,
• external interface types,
• external query types.
• Takes into account the complexity of each
  parameter. (Defined by the FP users group)
• Adjustment to programming language C 128
  lines of code for 1 FP.
• NEEDS adjustments to the development environment.
  (subjective, difficult to define)

7
Estimation Techniques (Cont.)

• Object Points Good for systems developed using
  High Level Application Building Tools (based on
  OBJECTS definition)
• Screens Reports object points based on number
  of tables accesses to generate the object
• 3GL component always weights as difficult
• Adjustment according to relative complexity of
  screens, reports and logic code (3GL). Reused
  objects. Productivity based on the developers
  experience and capabilities.

8
Estimation Techniques (Cont.)

• COCOMO II A Parametric Model
• Person MonthA X Size(SLOC) exp (SF) X EM
• A constant based on project type (organic,
  embedded, semi detached)
• SF 0.91 0.01 X SUM( EDR) (Exponent Driver
  Rating)
• EDR is based on existing/lack of
• similar cases in the past, development
  flexibility, architecture/risk resolution, team
  disperse, process maturity)
• EM EMa X EMb X . X EMk
• EMi Effort Multiplier(i) depends on the stage
  of the project when the evaluation is executed
  (Project Definition, Early Design, Post
  Architecture)
• Early stage samples schedule pressure,
  personnel experience
• Post architecture samples use of software tools,
  product complexity

9
Estimation Techniques (Cont.)

• Additional effort (time, cost) should allocate to
  requirements and specifications
• Additional effort should be added to cover
  management activities, overheads, suppliers,
  travels etc.)
• Additional effort should be added to
  implementation, training, cleanup activities.
• Conversion of data is an integral part of the
  project effort .

10
Risk Breakdown Structure (RBS) - Example
11
Risk Assessments

• Identification and Ranking (importance,
  likelihood)
• Risk and NPV ( add risk factor to the discount
  rate)
• Risk Profile Analysis (change a risk factor by
  /- 5 and recalculate the costs. Evaluate the
  sensitivity to the change. Focus on highly
  sensitive factors as possible risks)

12
Identification and Ranking (importance,
likelihood)

• Basic Risk Matrix Rank High, Medium, Low and
  Unlikely to the Importance and Likelihood of each
  risk.
• Deal (in the planning and decision stages) with
  risks that are ranked high on both importance and
  likelihood.

13
Reused Code

• Can reduce cost, time and increase quality.
• But. It has its own risks. (poor documentation,
  bad design, no testing procedures, etc.)
• Codes types New, Modifies, Reused, Legacy
• Simple adjustments to lines of code effort
• New 100
• Modifies 60
• Reused 30

14
Reused Code Example
15
Cost Benefit Analysis

• Business Plan
• ROI Return On Investment
• Risk Analysis
• Alternative Approaches
• Decision Trees Evaluate alternatives by
  multiply the probability of occurrence by
  expected cost to evaluate expected value.
• Build/Buy/Customize/Outsource/Enhance
• Cost Streams Definitions/Requirements, Solution
  Selection, Licenses, Hardware, Real Estate,
  Travel, Legal, Communication, Training,
  Development, Testing, Users Participation,
  Integration, Changes to Current
  Systems/Workflows, Maintenance, Ongoing-Support,
  Enhancements, Retirement.

16
Decision Tree(Market Expansion Market Growth
Beyond the Capability of the Enhanced System to
handle the growth)
decision
enhance
Npv10000
Npv40000
enhance
buy
Probability 0.2
0.8
0.2
0.8
market expansion
no expansion
market expansion
no expansion
-50000
Npv -100000
75000
250000