| Abstract |
: |
Various software developmental issues in pair programming environment such as,reducing defects, better bug fixing, optimumknowledge combinationofexpertise and thereby reducing developmental time etc.,have been reported in literature. They are however found mostlyoncomparative studies with single programmingofs/w projects.These characteristics arealso found to be dealt withmostly in isolation bythe available literature.Whether certain personality shortfalls among pair programmers in some of the above specified programmer characteristics would influence the efficiency in reducing s/w defects?Thisissue mightdemand forempirical social factors of programmer characteristics for further research purposes.We have reported in our earlier work that defect density (defective codes to overall code ratio) has reduced in small sized s/w projects, when domain experts have been paired with conventional programmers.This paper will be an extension of that, which attemptsfor an investigation on delimited personality shortfalls, namely talent in combination with domain job matching. The investigation aims at studying the influence of cross trainingmade on inexperienced programmers (of the pair) in cross domains with an objective to determine whether these shortfalls would be rectified and therebythe defect density of small sizeds/w development projects would decrease?Experimental setups, social surveys and the results reported in this paper form a part of a whole research program of the authors, and hence some of the relevant input data have beendrawn from our earlier published work [SunithaK. S &Nirmala, K, 2015]. When s/w defects relate to lines of coding (LOC) or in other words the developmental efforts, due to personality shortfalls might directly influencethe concerned paired humans. It is hypothesized that the domain experts (navigators) could controland influence the programmers (drivers) one anotheror in a combined fashion.The paper presents relational study results between defect densities and the talent personality shortfall in combination with job matching of 5 selected small sized s/w development casestudies. For the purpose of experiments,apart from the experts who would act as navigators, the driver programmers are selected from a control group of graduate students. For the experiments presented in this paper, the personality shortfalls are represented in qualitative metric forms and the results would provide an empirical cue to theinfluence of the defect densities.The results will demonstrate the major role played bythe cross training component on the non-domain programmers (the drivers). The results reported in this paper will be of importance to s/w engineering researchers and also provide utility values to s/w project managers. |
