Early Design Model : COCOMO-II

It is used at the Stage – II in COCOMO -II models and supports estimation in early design stage of project. Base equation used in COCOMO – II models is as follows –

PMnominal = A * (size)B 
where PMnominal = Effort for the project in person months
A = constant representing the nominal productivity where A = 2.5
B = Scale Factor
Size = size of the Software

The early design model uses Unadjusted Function Points (UFP) as measure of size. This model is used at the early stages of software project when there is not enough information available about size of product which has t be developed, nature of target platform and nature of employees to be involved in development of projector detailed specifications of process to be used. This model can be used in either Application Generator, System Integration, or Infrastructure Development Sector.

If B = 1.0, there is linear relationship between effort and size of product. If the value of B is not equal to 1, there will be non-linear relationship between size of product and effort. If B < 1.0, rate of increase of effort decreases as the size of product increases. If B > 1.0, rate of increase of effort increase as the size of product is increased.

This is due to growth of interpersonal communications and overheads due to growth of large system integration. Application composition model assumes value of B equal to 1. But Early Design Model assumes value of B to be greater than 1.

Thus here the basic assumption stands as effort on project usually increases faster than size of product. However, value of ‘B’ is computed on basis of scaling factors that may cause loss of productivity corresponding to an increase in the size as follows –

1. Precedentness :
   It reflects the experience on similar projects previously. This is applicable to individuals as well as organizations both in terms of expertise and experience. High value would imply that organization is quite familiar with application formula and very low value means no previous experience or expertise. The value for scale factor is 6.20(Very Low), 4.96(Low), 3.72(Average), 2.48(High), 1.25(Very High) and 0.00(Extra High).
2. Development Flexibility :
   It reflects degree of flexibility in development process. Low value would imply well defined process being used. Very high value would imply that the client offers very general idea of the product or project. Value for the scale factor is 5.07(Very Low), 4.05(Low), 3.04(Average), 2.03(High), 1.02(Very High) and 0.00(Extra High).
3. Architecture Risk and Resolution :
   It represents degree of risk analysis being carried out during course of project. Low value would indicate little analysis and very high value would represent complete and thorough risk analysis. Value for the scale factor is 7.07(Very Low), 5.65(Low), 4.24(Average), 2.83(High), 1.41(Very High) and 0.00(Extra High).
4. Team Cohesion :
   Reflects the team management skills of the employees developing the project. Very low value would imply very low interaction and hardly any relationship among the members however high value would imply great relationship and good team work. The value for scale factor is 5.48(Very Low), 4.38(Low), 3.29(Average), 2.19(High), 1.10(Very High) and 0.00(Extra High).
5. Process maturity :
   Reflects process maturity of organization. Very low value would imply organization has no level at all and high value would imply that organization is rated as highest level of the SEI-CMM. The value for scale factor is 7.80(Very Low), 6.24(Low), 4.68(Average), 3.12(High), 1.56(Very High) and 0.00(Extra High).

The value of B can be calculated as -
B = 0.91 + 0.01 * (Sum of rating scaling factors for project)

When all scaling factors of project are rated as extra high, best value of B is obtained which is equal to 0.91.

When all scaling factors are very low worst case values of B is obtained as 1.23.

Hence value of B varies from 0.91 to 1.23.