A systematic literature review on modified condition and decision coverage

Abstract

One of the requirements of Modified Condition and Decision Coverage (MCDC) is that test cases should be selected to demonstrate that a condition can independently affect the outcome of a decision. However, it is known that test cases cannot be selected to satisfy this "independently-affect" requirement when a decision contains strongly coupled conditions. Researchers have proposed various forms of MCDC in addressing this issue. In last 20 years, these variant MCDC forms have been discussed and applied as test data adequacy criteria in white-box testing, and as fault revealing techniques in empirical studies. In this paper, we performed a systematic literature review (SLR) on these MCDC variants. By clarifying the existing meanings of these variants, we are able to define another MCDC form. Moreover, we performed an empirical study to compare the fault detecting ability of this new form with other existing MCDC variants. The results showed that this new form of MCDC has a higher fault detecting ability than the others.

A new disjunctive literal insertion fault detection strategy in boolean specifications

“…They are extracted from two different avionic software systems and the Random Boolean expression set used in Yu et al 23 The first software system is traffic collision avoidance system II (TCAS‐II), an aircraft collision avoidance system, 24 and the second software system is line replaceable units (LRU) that belong to five airborne systems across two different airplane models 14 . Hence, the TCAS‐II and LRU expression sets are also used in other experiments 17,20 to detect different types of faults in the expressions. The three sets of expressions are referred to as TCAS‐II expressions : There are 20 expressions in this set as published in Weyuker et al 3 These expressions are in GF and are numbered from T01 g , T02 g , …, T20 g . LRU expressions : There are 20 expressions in this set; they were sampled from LRU.…”

“…Then, we have recorded their individual sizes and calculated the average size of each group of 1000 test suites. We have followed the experimental process as in our previous works 17,20 …”

“…Furthermore, researchers have performed a number of empirical studies to observe how effective the existing fault‐based testing strategies (e.g., branch operator (BOR) strategy, 4 RCDC, 11 MUTP strategy, 12 Cause‐effect graphs (CEG), 13 and variants of modified condition decision coverage (MCDC) 14 are in detecting LIFs in Boolean expressions 15‐17 . Though MCDC is a structural coverage criterion, researchers have been using MCDC definitions to generate test cases from a Boolean expression to detect faults in that expression 15,16,18‐20 . Furthermore, researchers have been working on improving the fault detection strength of MCDC using different strategies 21 .…”

Prioritizing MCDC test cases by spectral analysis of Boolean functions

“…For example, cause‐effect graphing is a popular requirement‐based testing introduced by Myers several decades ago and a recent study proposes the application of MCDC to generate test cases from these graphs . For further information on MCDC analysis, there exist some fundamental studies , including a systematic literature overview .…”

Moving Switching Functions to Continuous Domain

“…There are several forms of MCDC discussed in the literature [9], [10], [11], [12]. In this work, we employ …”