Cocomo II A Worked Example EEE493 2001

1
Cocomo IIA Worked ExampleEEE493 2001
Royal Military College of Canada Electrical and
Computer Engineering
Major Ron Smith smith-r_at_rmc.ca 1-613-541-6000
ext. 6030

• Major Greg Phillips
• greg.phillips_at_rmc.ca
• 1-613-541-6000 ext. 6190

2
COCOMO II -Exercise Scenario

• use COCOMO-II model to estimate the effort and
  schedule for the Personal Income Tax (PIT)
  software application
• approximately 648 unadjusted function points
• given that requirements prototyping and a basic
  architectural design have been completed, use the
  Post-Architectural model
• PMNS a x Size b x ? EMi (i 1 to 16)
• certain scale and effort multipliers can be
  attained from the scenario description, others
  must be assumed to be nominal

3
Scaling Factors
?
?
?
?
?
4
Scaling Factors

• cause an exponential cost increase

5
Effort Multipliers (Post-Architecture)
every 1 mo.
6
Effort Multipliers (Post-Architecture)
?
?
?
?
?
7
Effort Multipliers (Post-Architecture)

• Personnel factors

8
Effort Multipliers (Post-Architecture)

• Project factors

9
Calculating Estimated Effort

• PMNS a x Size b x ? EMi (i 1 to 16)
• where a 2.94 (calibrated from 161 projects)
• b 1.01 0.01 x S SFj (j 1 to 5)
• 1.01 0.01 x (11343) 1.13
• ? EMi ACAP x APEX x PCON x SITE
• 1.22 x 0.89 x 1.17 x 0.84 1.067 Size
  648 UFPs 23 Java SLOC/UFP
• 14.9 kSLOC
• therefore
• PMNS 2.94 x 14.9(1.13) x 1.067
• 66.4 person-months

10
Sensitivity Analysis (1)

• compare the effect that documentation
  requirements might have upon the project effort
• for minimal doc requirements (DOCU 0.89)
• ? EMi ACAP x APEX x PCON x SITE x DOCU
• 1.067 x 0.89 0.950
• or more simply,
• PMNS 66.4 0.89 59 person-months
• and for excessive doc requirements (DOCU 1.13)
• PMNS 66.4 1.13 75 person-months

11
Sensitivity Analysis (2)

• what savings might be achieved if you were to
  change contracts and bring in a Level IV company?
  what else must you consider?
• now b 1.01 0.01 x (11341) 1.11
• and
• PMNS 2.94 x 14.9(1.11) x 1.067
• 62.9 person-months
• a possible savings of 5, but what about other
  factors such as application experience, ...

12
Reuse Analysis

• the question is will the effort involved in
  reusing the CTRC be sufficiently less than the
  effort to develop the same functionality from
  scratch during the PIT development?
• the CTRC can satisfy 20 of the required
  functionality for PIT (or about 130 UFPs)
• the CTRC consists of 4000 source lines of Java
• 130 UFPs represents approximately 3.0 kSLOC
  (Java)
• using the COCOMO II Reuse model, determine the
  reuse size estimate
• compare this reuse size to the expected
  development size of code
• is SizeRU lt SizeD ?

13
Reuse Software Guidelines
14
Determining Reuse Size

• SizeRU SizeA AA AAF (SU x UNFM) /100
• where
• SizeA 4.0 kSLOC
• and
• AA - of assessment and assimilation
• AAA - adaptation adjustment factor ()
• SU - software understanding increment (10 -
  50)
• UNFM - programmer unfamiliarity (0.0 - 1.0)

15
Adaptation Adjustment Factor

• recall, default efforts for each phase are
  assumed to be 40, 30 and 30 respectively
• AAF 0.4 DM 0.3 CM 0.3 IM
• 0.4(20) 0.3(30) 0.3(25)
• 24.5
• note that,
• IM - is the incremental of effort required to
  integrate the adapted software into the product
  as compared to integrating a newly developed
  equivalent sized piece of software

16
Assesment and Assimilation Increment
in percentage ()
17
Software Understanding Increment
?
?
?
18
Programmer Unfamiliarity Factor
19
Is Reuse Warranted ?

• SizeRU SizeA AA AAF (SU x UNFM) /100
• 4.0 4 24.5 20(0.8)/100
• 1.78 kSLOC
• therefore the reuse savings is 1.78 kSLOC versus
  3.0 kSLOC or 40
• accordingly reuse does appear to provide
  meaningful savings
• other issues to consider?
• how accurate are the redesign and recode
  estimates given the unfamiliarity of the
  programmers with CTRC

20
Calculating Estimated Schedule

• next class

21
Supplemental References

• Boehm, Barry, et al., Software Cost Estimation
  with COCOMO II , Prentice-Hall, 2000. ISBN
  0-13-026692-2.

22
Next ClassEstimating Labour Distribution