constructive cost model (COCOMO)
COCOMO (COnstructive COst Model) is an effort and cost estimation technique for the development of software. It is an algorithmic cost model that uses mathematical functions to relate specific software metrics to the cost of the software project, allowing for better planning and realization of quality attributes such as cost and schedule adherence of a project.
COCOMO was presented for the first time 1981 and revised in the context of COCOMO II 1995 substantially, so that also increasingly the special requirements of object-oriented software are considered. The method works with so-called Object Points. Apart from the function point method COCOMO is the most widely used software metric, which pursues however a differentiated approach. COCOMO is based essentially on an estimate of the program lines, Lines of Code( LOC) of the software product to be developed in Kilo Source Lines of Code (KSLOC).
The graded COCOMO method
This software metric is a staged procedure based on the Application Composition, Early Design and Post Architecture stages.
In the Application Comp osition stage, development effort is determined in employee-months( MM) based on an estimate of weighted Object Points divided by a standard value for employee productivity. The number of Object Points is determined by first counting the number of separately displayed screens and then weighting them according to their level of complexity. Additional Object Points are then added according to the programming language used. The number of object points determined in this way is weighted with a factor that expresses the reusability of the software and divided by the standard value for productivity.
In the Early Design stage, the effort is determined using the following algorithmic formula:
Effort (MM) = 2.5 x Production Size (exp B) x M
Here, the product size represents the number of KSLOC. This value is determined by determining Function Points and converting them using a table. As is usual with algorithmic estimation methods, the exponent B reflects the development effort that increases disproportionately with the size of the project.
B can take on values between 1.1 and 1.24 depending on the subjective influencing factors such as the quality of the team, the degree of innovation, flexibility and the security management of the project.
Factor M defines the project and process cost drivers such as Product Reliability, Reliability and Complexity (RCPX), Required Reuse, Reuse (RUSE), Platform Complexity, Platform Difficulty (PDIF), Personnel Experience, Personnel Experience (PREX), Schedule Requirements, Required Schedule (SCED), and Support Infrastructure, Team Support Facilities (FCIL). Each of these factors can take values between one (very low) and six (very high). This results in:
M = PERS x RCPS x RUSE x PDIF x PREX x FCIL x SCED.
In the Post Architecture stage, a similar formula is also used as in the second stage. However, the factor M is composed of an extended number of influencing factors. These can be classified into four different classes: In the product characteristics, the hardware platform, the personnel and the project characteristics.
The value of the product size is determined in the unit KSLOC according to the following formula 1:
Defined therein
- ASLOC for the potentially reusable Lines of Codes,
- AA for the test effort from 0 to 8,
- SU for software quality from 10 (well structured) to 50 (complex but unstructured),
- DM the percentage of design to be modified for reuse,
- CM the percentage of reusable program code, and
- IM the percentage of the total integration effort for reusable code.