Page 32 - DCAP304_DCAP515_SOFTWARE_PROJECT_MANAGEMENT
P. 32

Software Project Management




                    Notes              inputs and outputs, user interactions, external interfaces and files used by the system. The
                                       function points are then converted to  number of lines of  source code using the  tables
                                       provided by the COCOMO model.

                                   Application Area

                                   COCOMO  is  a  well-known experiential  algorithmic  cost  estimation procedure. It  is  well-
                                   documented, in the public domain and is supported by public domain and commercial tools.
                                   It has been extensively used and has a long pedigree from its first instantiation in 1981. The
                                   application of the first instantiation of the model was limited due to the quite large constraints
                                   on the development process. This issue has been mitigated by continued improvements on, and
                                   extensions  of the  model, resulting  in COCOMO  2. A  refinement  of the model for the  Ada
                                   programming language is available as well.
                                   Advantages


                                      Though it’s hard to pinpoint the exact cost of any given project, one can still obtain usable
                                       data by calculating optimistic and pessimistic estimates.

                                      Implementation and execution of the model is very easy and proficient. As a result, it is
                                       supported by public as well as commercial tools.
                                      COCOMO is a well-known and well-documented method.

                                   Disadvantages

                                      It is quite difficult to come up with acceptable estimates for the size of a project when the
                                       latter still in an early stage of development.
                                      The use of the number of lines of source code as a measure of difficulty is highly disputable.
                                       Even though COCOMO tries to take this into account by providing different tables for all
                                       major programming languages, there are still lots of discrepancy such as: expressivity
                                       differences between programmers, usage of subroutines, general code reuse, and etcetera.

                                      Some input parameters in the model cannot by determined quantitatively; they need be
                                       estimated as well. A few examples: experience  and productivity of the programmers,
                                       maturity and capability  of CASE  tools. The  accuracy of  the ultimate  estimates of  the
                                       COCOMO model depends considerably on the exactness of the initial ones.
                                      The COCOMO model has not been revised since 1995. Consequently, it is likely the model
                                       fails to take into account new theories and practices in  the Software Engineering field,
                                       resulting in worse estimates.

                                   Usage in Methodologies

                                   As confirmed above, the COCOMO model can only be applied when the project in question
                                   satisfies a given number of criteria. Moreover, it is advisable to try out other estimation techniques,
                                   as to get a feeling of the accuracy of the estimates  that have been obtained. Other probable
                                   techniques include:
                                      Expert judgment
                                      Estimation by analogy

                                      Other algorithmic cost estimation models





          26                                LOVELY PROFESSIONAL UNIVERSITY
   27   28   29   30   31   32   33   34   35   36   37