Such rule satisfies demand for metric that they provide initial estimates of size, cost, and/or effort of to be developed software, being thus, sufficiently used in the market. For Smith (1999) apud Pressman (2006), before using cases of use for estimate of cost/effort of software, must be established some prerequisite ones, as: level of structural hierarchy; average size (numbers of pages) of each use case; type of software (software of real time, business, scientific engineering/, inlaid); an sketch of the architecture of the system. After the definition of these prerequisite ones, ' ' empirical data can be used to establish the esteem number of LOC or FP for use case. Historical data are then used to calculate the effort necessary to develop sistema' '. (PRESSMAN, 2006).
2,2 EMPIRICAL MODELS According to Pagno (2010), this model of estimate empirically uses formulas derived from data of an limited sample of projects with the objective to foretell information of planning of project. They can be used as complement to the decomposition techniques, offering important estimates through its proper method. By being derived from a sample of concluded projects already, the empirical models are not adjusted to all the software classrooms and all the development environments. Thus, the estimate model must be calibrated, in order to reflect the work conditions local. Pressman (2006), says that the model must be tested applying the collected data of finished projects already and be compared with the real results. If the discrepancy will be great, the model must be modified and retestado, before it can be used. The empirical models are derived by analysis of regression of collected data of previous projects of software. 2.2.1COCOMO model COCOMO (Constructive Cost Model) is an empirical model that it aims at to measure effort, stated period, size of necessary team and cost for the development of a software project, since that if has the dimension of project.