Presentation is loading. Please wait.

Presentation is loading. Please wait.

Briand4 1 Property-Based Software Engineering Measurement Lionel Briand, Sandro Morasca, Victor Basili IEEE TOSE Jan 96.

Similar presentations


Presentation on theme: "Briand4 1 Property-Based Software Engineering Measurement Lionel Briand, Sandro Morasca, Victor Basili IEEE TOSE Jan 96."— Presentation transcript:

1 briand4 1 Property-Based Software Engineering Measurement Lionel Briand, Sandro Morasca, Victor Basili IEEE TOSE Jan 96

2 briand4 2 Basic Assumptions u Properties can determine the type of measurement

3 briand4 3 Systems and Modules – p70 u A system,S, is a pair where E is the set of elements and R is the set of relations between the elements u A system m = is a module of S iff Em is a subset of E and Rm is a subset of Em cross Em and Rm is a subset of R.

4 briand4 4 Size u PS1- Nonnegativity u PS2- Null value u PS3- Module additivity

5 briand4 5 Is LOC a size measure? u What is E and R? u What is Size(S)? u Are all the properties satisfied –Nonnegativity –Null value –Module additivity

6 briand4 6 Which of the following are size measures? Why or why not?  Halstead’s vocabulary,  u Halstead’s length, N u The constant function zero

7 briand4 7 Can you think of u A size metric that does not satisfy the properties? u A non-size metric that does satisfy the properties?

8 briand4 8 Length u PL1 – Nonnegativity u PL2 – Null value u PL3 – Nonincreasing monotonicity u PL4 – Nondecreasing monotonicity u PL5 – Disjoint modules

9 briand4 9 Length u Can you think of a length measure other than those mentioned in the paper? (nesting depth and DIT (depth of inheritance tree).

10 briand4 10 Length u “Properties L1 through L5 hold when applying the admissible transformation of the ratio scale. Therefore, there is no contradiction between our concept of length and the definition of length measures on a ratio scale.” u What does this mean?

11 briand4 11 Complexity u PC1 – nonnegativity u PC2 – Null value u PC3 – Symmetry u PC4 – Module monotonicity u PC5 – Disjoint Module Additivity

12 briand4 12 Cohesion u PC1 – nonnegativity and normalization u PC2 – null value u PC3 – monotonicity u PC4 – cohesive modules

13 briand4 13 Coupling u PC1 – nonnegativity u PC2 – Null value u PC3 – monotonicity u PC4 – Merging of modules u PC5 – Disjoint Module Additivity

14 briand4 14 Measurement Theory u How does this theory match measurement theory?


Download ppt "Briand4 1 Property-Based Software Engineering Measurement Lionel Briand, Sandro Morasca, Victor Basili IEEE TOSE Jan 96."

Similar presentations


Ads by Google