1. Use of Estimation Methods on finished projects. Comparative Evaluation between the theorical estimation and the real time of development. Alejandro Fernández Miguel Díaz Cristina Ramos Gabriel Tena

2. Content  Deviation meaning.  Projects Introduction.  Estimation Methods: Function Points and COCOMO.  Results of the theorical estimation methods.  Comparative with real time: Deviation results.  Conclusions.

3. Deviation meaning.  Difference between estimation and real results. Estimation: 8Real: 4 Deviation

4. SICODI  Application to control diabetes’ parameters of patients.  Options in order to follow the illness pattern by the doctor.  Simple GUI: destinated to all kind of potencial users.  Multiuser suport.  Reports generation chosen by users.  Reports view suport.

5. SICODI (II)  Divided into 4 Software Modules: - Users. - Graphics and Reports. - Measures. - Importing/Exporting.  Developed in C#, using a Db4o by one developer with no previous knowledge about them.

6. Qweld (I)  Application for the visualization and monitoring of the weld process.  Main Inputs: Weld Files  Main functions: Graphical and textual representation of the weld parameters  Main Outputs: Reports about the visualizated weld.

7. QWeld (II)  Structure of Qweld: Modules  Connection.  Welds’ management.  Parameters representation.  Reporting.  Development: Four generation language without any experience in it.

8. VEPV (I)  Application for the management of the elechtronic votes.  Posibility of change the votation rules and configuration.  Simple access using a web interface.  Security Warranty: it has to protect the confidential data.  Counting votes facilities.

9. VEPV (II)  Two roles: - User:  Vote register.  Vote consulting. - Administrator:  Creates the votation rules and information.  Configure voting users.  Results consulting.

10. ESTIMATION METHODS  Objective  Measure the effort required so as to calculate time/cost of developement.  Different kinds of methods: - Objective Methods. - Subjective Methods.  Deviation calculate methods  FP & COCOMO.

11. FUNCTION POINTS  Used to… Measure what the user wants and what the user gets Measure the technology used to deploy the system separately Provide a size metric that supports quality analysis and productivity Provide a measure for estimating software Provide a normalization factor for comparing different pieces of software.

12. FUNCTION POINTS (II)  Calculated with… External Input (EI) External Output (EO) External Query (EQ) Internal Logic File (ILF) External Interface File (EIF).

13. COCOMO  Used to… Develop a model for software development time and cost estimation that conformed to the lifecycles used in the 1990s and the first decade of 2000. Develop software cost databases and tool support capabilities for continuous model improvement. Provide a quantitative analytical framework and set of tools and techniques for evaluating the effects of software technology improvement on software lifecycle costs and schedules.

14. FUNCTION POINTS SICODI

15. COCOMO SICODI

16. COMPARISON SICODI Real dataEstimated dataDeviation HoursEeffortEffort% Plans and requirements 2801,8420,643-65,09% Product Design4222,7761,562-43,73% Programming8505,5925,7923,58% Integration and test 1150,7571,83558,74% Total166710,9679,832-10,35%

17. FUNCTION POINTS Qweld

18. COCOMO Qweld

19. COMPARISON Qweld Real dataEstimated dataDeviation HoursEeffortEffort% Plans and requirements 1721,1320,526-53,53% Product Design2121,3951,277-8,46% Programming4963,2634,78146,52% Integration and test 840,5531,45461,96% Total9646,3438,03826,72%

20. FUNCTION POINTS VEPV

21. COCOMO VEPV

22. COMPARISON VEPV Real dataEstimated dataDeviation HoursEeffortEffort% Plans and requirements 1801,1840,585-50,59% Product Design1801,1841,42120,02% Programming4863,1975,20362,75% Integration and test 840,5531,73368,09% Total11707,6978,0384,24%

23. COMPARISON

24. LOCAL CALIBRATION  Best fit to your environment  Calibrated to your completed projects  New Effort or Schedule Equation PM nominal=A x (Size ) B - Multiplier Only - Multiplier and Exponent Base  Not necessary to add a new Cost Driver (but it is possible)

25. LOCAL CALIBRATION Justification Fixes some shortcomings of COCOMO:  Unusual standard for counting effort  Hours per Staff-Month  What effort is included  Nonstandard cost driver definition  Local definition of PCAP High, etc.  How SLOC are counted  Physical lines

26. LOCAL CALIBRATION Justification  Accuracy. A local calibration means you get an equation that is the best fit to YOUR environment.  You may want to do several local calibrations – for different types of projects, or different tool sets.

27. LOCAL CALIBRATION Required data  Actual SLOC  EAF & SCED setting, Scale Factors  Actual Effort, Hours per Staff-Month  Actual Duration You need:  5 data points to calibrate the multiplicative constant  10 data points to calibrate the constant & exponent

28. LOCAL CALIBRATION Some considerations  To combine data points in a calibration, they must be comparable  Hours per Person-Month  Must calibrate with Normalized Effort Normalized Effort = Actual Effort*(HPM/152)  SCED Cost Driver  Must calibrate with Normalized Duration Normalized Duration = Actual Duration / Schedule Acceleration

29. CALIBRATION TOOLS  USC Tool  Calico  Spreadsheet  http://www.softstarsystems.com/  Homegrown spreadsheet or program

30. CONCLUSIONS  Local calibration is good for enterprises.  Start collecting data now.  As few as 5 data point will get you started.  Multiplier-only calibration is most conservative.  Use a good tool to calibrate COCOMO.

31. Use of Estimation Methods on finished projects. Comparative Evaluation between the theorical estimation and the real time of development. Alejandro Fernández Miguel Díaz Cristina Ramos Gabriel Tena