Defining and validating measures for object-based high-level design
read more
Citations
Research methods in social relations
A unified framework for coupling measurement in object-oriented systems
A unified framework for coupling measurement in object-oriented systems
Software Module Clustering as a Multi-Objective Search Problem
References
Applied Logistic Regression
Applied Logistic Regression.
A metrics suite for object oriented design
Object-Oriented Software Construction
Related Papers (5)
Frequently Asked Questions (11)
Q2. What are the future works mentioned in the paper "Defining and validating measures for object-based high-level design" ?
Their future work will be four-fold to: 1 ) analyze more systems. 3 ) further validate and refine the measures the authors defined in this paper. 4 ) be consistent with their current objectives, the authors will address the issues related to building measure-based empirical models earlier in the life cycle.
Q3. What are the main strategies to validate software measurement hypotheses?
In order to validate software measurement hypotheses empirically, one can adopt two main strategies:1) small-scale controlled experiments, 2) real-scale industrial case studies.
Q4. Why are higher module imports seen in TONS?
Higher module imports may be due to an increase in complexity over time of the systems developed in the studied environment or to the difference in application domain.
Q5. What is the definition of a high-level design of a software system?
The high-level design of a software system is a collection of module and subroutine interfaces related to each other by means of USES and IS_COMPONENT_OF relationships.
Q6. What other methods could be used to define the cohesion of software parts?
other definitions and modeling techniques could be used, e.g., number of faults and least-square regression, respectively.
Q7. What are the early indicators for fault-prone software?
In particular, the authors have identified some early indicators for fault-prone software that may be interpreted as cohesion and coupling measures.
Q8. What is the way to measure the impact of high-level design measures across projects?
In a given application domain, the impact of the defined high-level design measures seems to be relatively stable across projects.
Q9. What are the basic relationships between data declarations and subroutines?
The DD-interaction relationships can be defined in terms of the basic relationships between data declarations allowed by the language, which represent direct DDinteractions, i.e., not obtained by virtue of the transitivity of interaction relationships.
Q10. What are the important criteria for coupling?
Measures differ according to several criteria and the most important ones are: the types of connection/ dependency contributing to coupling, the locus of impact, i.e., import vs. export coupling, the domain of the measure, its level of granularity, i.e., how connections are counted, and,as for cohesion, how indirect connections and inheritance are handled.
Q11. What is the lnlikelihood of a model without variables?
R2 is defined by the following ratio:R LL LLLL 2 S S =-whereLL is the loglikelihood obtained by Maximum Likelihood Estimation of the model described in formula (2) LLS is the loglikelihood obtained by Maximum Likelihood Estimation of a model without any variables, i.e., with only C0.