Function Point Methods and Their Limitations

In 1979 Albrecht developed a conception for function points. This function points come under the metric called application size (Kitchenham, 1997). Through this function point any effort present in the application can be detected.

Five main kinds of fundamental elements are present in function points, which include External Interface File (EIF), External Inquiry (El), Internal Logical File (ILF), External Output (EO), and External Input (El). For every element, the complexity is classed as high, average, or low.

The Complexity Adjustment Factors (CAF) and Unadjusted Function Points (UFP) are multiplied to obtain the function point. The equation for this function point is represented as follows.

The summation of factor of complexities (Wij) multiplied with number of classified elements (Iij) within the application is referred to as UFP. The equation for this UFP is represented as follows:

TABLE 2.2 Weights of /,

FUNCTION UNITS

AVG

LOW

HIGH

Internal Logical Files

10

7

15

External Output

5

4

7

External Interface File

7

5

10

External Input

4

3

6

External Inquires

4

3

6

Where W0 is the weight of /,, and Table 2.2 represents the values of W4

For the entire application the process complexity and environment was evaluated through CAF. The features of 14 basic systems have some influence and this is rated on a scale between 0 and 5 (5 = Essential, 4 = Significant, 3 = Average, 2 = Moderate, and 1 = Incidental). This rating was given based on its effect in the project.

Then CAF is computed using following formula:

CAF = 0.65 + .01 where I = 1 to 14

Example 1:

A project is considered along with its subsequent functional units:

  • • Number of external outputs = 4.
  • • Number of external inputs = 5.
  • • Number of external interfaces = 2.
  • • Number of external enquiries = 3.
  • • Number of Internal files = 2.

It is assumed that all the weighting factors of elements are simple additional complexity adjustment factors and were found to be average. For the given project a function point is evaluated.

SOLUTION

Function Point Elements

Number of Elements

Weights of Element

Wij x /fj

Number of external inputs

5

3

15

Number of external outputs

4

4

16

Number of external enquiries

3

3

9

Number of external interfaces

2

5

10

Number of Internal files

2

7

14

UFP =

64

The drawbacks and complications that were found while utilizing the function point method is explained by Kitchenham in his article. In the previous section it was explained that the complexity of the element was classified as high, average, and low. This implies that the measurement was using an ordinal scale and the use of absolute scale counts was declined. Basically, the measurement that is obtained from an ordinal scale is impossible to add together because the elements such as high, average, and low are separate matters. In case of utilizing similar factors, adding cannot be achieved. In an ordinal scale the measurement of size can be achieved.

It can say that the size of one particular system is bigger when compared to another but the value for the size cannot be found through this function point approach. The number of elements to be considered will be decided through examining the Symons Mark function point construction. It was found that for equating the entity, output, and input no standard conversion factors exist (Kitchenham, 1997).

The correlation between the elements in the function point must be found. This is considered a significant criterion. The function point is viewed to be inconsistent, which is another limitation.

 
Source
< Prev   CONTENTS   Source   Next >