CSCI 510 Midterm 1, Fall 2017

Transcription

1 CSCI 510 Midterm 1, Fall 2017 Monday, September 25, questions, 100 points If registered DEN student, please circle: Yes Last Name: First Name: USC ID: Question 1 (30) Question 2 (40) Question 3 (30) Total (100) When working exam questions, please assume that non-simple estimating situations do not occur in the problem, unless they are mentioned explicitly. Specifically, unless mentioned otherwise in the problem statement: The Post-Architecture model is used (not the Early Design model); There is no Adapted (reused) software (i.e., all software is new); Requirements Volatility (REVL) is zero; There is no automatic translation; The Inception and Transition phases are not part of the estimate (i.e., the model equations give the effort and schedule directly); No risk reserves are needed; The nominal schedule is to be used (i.e., SCED% = 100%); and Only one module is to be produced. As a related shortcut, a problem statement will frequently say something like "all other Effort Multipliers should be taken to be Nominal". As you have learned, the numeric value for any Effort Multiplier at the Nominal level is 1.0. As another shortcut, there is a standard approximation for Scale Factors: When all Scale Factors are Nominal, you should use 1.1 as the value for the exponent E. Use at least 3 significant digits for your final answer (and your intermediate results). More significant digits is acceptable. (To avoid some rounding difficulties, 4+ significant digits is recommended.)

2 1. Business Case Analysis, 30 points Solution, Rubric (Graded by Jim) As indicated in the ICSM book Table 6-4, the MedFRS Initial Operational Capability (IOC) will cost $2 million to develop. It will also cost $0.5 million to transition into use and $1.2 million per year to operate and maintain. Besides the benefits in human life and health saved, which are generally agreed not to be given dollar values, the conservative annual savings in hospital and operations costs are estimated to be 20% for each of the 4 areas in the Ensayo region. Each of the 4 areas averages $9 million per year in hospital and operations costs. Compute the Return on Investment (ROI) for the IOC net savings (IOC benefits to date minus IOC costs to date) produced for the Ensayo region by the IOC investment at the end of Years 1, 2, and 3 of the IOC operation. The IOC development and transition is complete by the beginning of Year 1. IOC Annual savings per area = $9M/area * 20% = $1.8M. Total IOC annual savings per year = 4 * $1.8M = $7.2M (4 points: 1 point for including all 4 areas; 2 points for otherwise correct form (could start by multiplying $9M by 4); 1 point for correct answer.) IOC savings to date after Years 1,2,3 = $7.2M * Number of years in operation (4 points: either for explicit formula, or for appropriate figures in all three year calculations.) IOC cost to date: $2.5M + $1.2M * Number of years in operation (4 points: either for explicit formula, or for appropriate figures in all three year calculations.) Year 1 ROI = ($7.2M - $3.7M) / $3.7M = ($3.5M / $3.7M) = (6 points: 2 for correct equation form, 1 for correct savings amount, 1 for correct cost amount, 2 for correct final answer.) Year 2 ROI = ($14.4M - $4.9M) / $4.9M = ($9.5M / $4.9M) = 1.94 (6 points (same as Year 1).) Year 3 ROI = ($21.6M - $6.1M) / $6.1M = ($15.5M / $6.1M) = 2.54 (6 points (same as Year 1).) 2

3 2. COCOMO II Estimate, 40 points The MedFRS long-range plan is to devote the first year of its IOC operation to analyze stakeholder feedback and operational experience, and to use the results to prioritize the features and quality attributes to be developed in the next 2 years and $2.5 million budget for the MedFRS Full Operational Capability (FOC). 2.1 FOC COCOMO II Estimate, 30 points. The FOC estimated size is 70 KSLOC. Its scale factor and cost driver ratings are all Nominal except for High ratings for the scale factors Team Cohesion (TEAM), Architecture and Risk Resolution (RESL) and Process maturity (PMAT), and the following cost driver ratings: Very High for RELY, and High for DATA, CPLX, ACAP, and PCAP. The cost per person-month is $8K. Estimate the FOC project effort, cost and development schedule. Answers show lots of significant digits; only 3 are required. Scale Factor = * ( ) = (7 points total 1 for correct equation form, 0.5 for constants (0.91, 0.01), 1 point each for correct TEAM (2.19), RESL (2.83), and PMAT (3.12), 0.5 point total for correct nominal ratings for other scale drivers (3.72, 3.04), and 2 points for correct final exponent value) EAF = 1.26 * 1.14 * 1.17 * 0.85 * 0.88 = (5 points total 0.5 for RELY, 0.5 for DATA, 0.5 for CPLX, 0.5 for ACAP, 0.5 for PCAP, 2.5 for correct EAF value) Effort = 2.94 * * = 2.94 * * = PM. (6 points total 2 for equation form, 1 for correct constant (2.94), 1 for correct KSLOC, 2 for correct PM) Cost = $8K * = $ K (3 points total 1 for $8K, 1 for using Effort result, 1 for final cost) Schedule Exponent = * ( ) = (5 points total 2 for equation form, 1 for constants, 2 for correct schedule exponent) Schedule = 3.67 * (333) = 3.67 * = months (4 points total 1 for equation, 1 for constant, 2 for correct schedule) 3

4 2.2 Fitting Budget and Schedule, 10 points. Do the estimated cost and schedule fit within the planned and budgeted $2.5 million and 24 months? Where they do not? Identify at least 3 workable strategies for modifying the estimate inputs to get the cost and schedule to fit. (Note: Recalculations not necessary). An example of a workable strategy is: Convince the stakeholders that not as much reliability is needed, so that RELY can be reduced from Very High to High. Of course, this strategy MAY NOT be used as an answer (additionally, this strategy would not be feasible from a safety standpoint). Clarifying note: (A workable strategy should state the change to be made and explain how the change would affect COCOMO parameter values.) Schedule fits within 24 months. $ K Cost is above $2.5M. (1 point total: 0.5 points for each correct comparison) Workable strategies for modifying the estimate inputs to get the cost and schedule to fit include (9 points total: 3 points per good strategy 1 point for change/potential solution, 2 points for correct and good strategy/explanation to reach the solution): Examples of workable strategies: Dropping low-priority features to reduce size Providing developer completion bonuses to improve PCON rating Finding more experienced performers to improve APEX, LTEX, PLEX ratings Collocating the development team to improve the SITE rating. Using key members of the IOC team to increase the PREC rating Reusing significant parts of the IOC to reduce ESLOC Hire teams that have worked together before, or have some non-expensive teambuilding activities to increase TEAM rating 4

5 3. True-False Questions. 30 points; 3 points each. Put either T or F on the line preceding the statement. Grading will be based on this alone, not any additional statements or explanations provided. 1. F Agile methodologies are not compatible with ICSM, since it is a risk-driven process model generator, with specific life-cycle phases. (ICSM, pages 22-23) 2. F When doing a business case, the common denominator is schedule. (EC-6, slide 6) 3. T When determining whether or not to reuse a software component, a good approach is to use the COCOMO II reuse model to determine the ESLOC for the reuse effort for the component, and compare that figure to the component's actual number of lines of code. (EC-5, slide 17) 4. F According to the Risk-Driven Scalable Spiral Model, it's not possible to have both rapid response to change and high assurance. (EC-1 slide 19) 5. T One estimation challenge, according to the Dual Cone of Uncertainty, is that the longer the development cycle, the more likely it will be that uncertainties in competition and technology evolution and changes in organization and mission priorities can make the originally-defined system obsolete. (EC-7, slide 9) 6. T Finding a sweet spot by balancing amount of testing and time to ship is an example of applying ICSM Meta-Principle 4+ (Risk-Balancing). (EC-2, slides 19, 23) 7. T "Requirements understanding" is an effort multiplier (cost driver) for COSYSMO. (EC-4, slide 15) 8. F It is low-risk to ask a project to maximize, minimize, or optimize a system's aspect, because such problems are a normal part of software development. (ICSM pages 40-41) 9. T Part of the COCOMO II model is the exponent, which handles the phenomenon of "diseconomy of scale". (EC-3, slide 7) 10. T Part of the COCOMO II model is the exponent, which handles the phenomenon of "economy of scale". (EC-3, slide 7) 5