One of the technique which measures the scale of the software. It is utilized in estimate of development man-hour. With "the function" of the software on the basis of, those where it keeps acquiring the score, function point from complexity and the like of the processing contents, the point of all functions of the software total scale and man-hour deduce.
Basis of calculation method is as follows. First, the case where it counts function and the like it makes clear type it counts to somewhere. Next, input, external output, external collation and the internal logical file outside it is called user function type concerning the system which becomes the object of appraisal, it counts the number of external interfaces as the number of functions. Whether the function which extent referring to which file, being something which counts whether whether it exchanges, before making the program, it can estimate from DFD and ERD etc. of stage.
And, 3 stages (easy normally complicated) with appraising the respective degree of hardness, score conversion (point conversion) doing, putting the following factor on that point, it designates the value which it sums up as reference level.
Easy Normal Complicated
External input 3 4 6
External output 4 5 7
External collation 3 4 6
Internal logical file 7 10 15
External interface 5 7 10
- Next, with item below you appraise the complexity at 6 stages 0 - 5 in regard to system characteristic, total that and calculate adjustment value.
- Complexity of processing Plural use with sights Readiness of modification
- Data communication Transaction quantity System operability
- Online renewal Efficiency condition Reuse characteristic
- Online renewal High load and constitution Installation readiness
- Distributed data processing Performance of end user
The on the basis of reference level and the adjustment value which you deduce, the function point is calculated with formula below.
Function point = reference level × (0.65 + adjustment values/100)
As the value which estimates system scale, other than the number of functions "the number of lines of the source code (SLOC: Source lines of code)"and there is a" file size "and so on, but not be able to calculate these, that it is not the stage where system important matter almost became firm in addition because it depends on enviroment for software development and operational environment, as objective index and there was a fault, which it handles ? leprosy.
There is a feature that as for function point method, before entering into programming phase, user important matter becomes settled, it is possible to estimate system scale at the stage where necessary function is visible. Before starting project generally, you estimate roughly at stage with function point method, it is being at the point where specification becomes firm, doing again to measure for the second time, the technique which increases the estimate precision of system scale is used. In addition, it can utilize in index of the kana every investment decision making which purchases whether it compares the scale of the system which differs at the time of system supply, develops anew.
The allene J Al blurring person of IBM (A.J.Albrecht) the person proposed function point method, in 1979, after that could add improvement in the user group of IBM. The United States IFPUG (International Function Point Users Group) you establish in 1986, assure international spread fixation. In US the indian function user meeting (JFPUG) is established in 1994. In addition, with ISO/IEC, standardization job of "functional scale estimate technique" is advanced function point method to the base. |
