Finding Self-organized Criticality in Collaborative Work via Repository Mining

Merelo, J. J.; Castillo, Pedro; García-Valdéz, Mario

doi:10.1007/978-3-319-59153-7_42

Cited by 1 publication

(3 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…These studies can be found in [Wu et al, 2007,Gorshenev andPismak, 2004], but they are not the only ones of this kind. Tests on the possible self-organized criticality have also been sought in another type of repositories and projects, which are enumerable in size and purpose [Merelo, 2016,Merelo-Guervós, 2016a,Merelo-Guervós, 2016b, Merelo et al, 2017a.…”

Section: State Of the Artmentioning

confidence: 99%

“…The general trend being the lack of statistical evidence that supports the hypothesis of the existence of a power law, the truth is that as described in [Newman, 2005,Clauset et al, 2009 the most common method for making subsequent adjustments, Leas Square Medium [Merelo et al, 2017a,Arafat and Riehle, 2009, Merelo-Guervós, 2016b, usually implies the existence of a greater error than the possible alternatives.…”

Section: State Of the Artmentioning

confidence: 99%

“…Following the approach outlined in [Merelo-Guervós, 2016b, Merelo, 2016, Merelo et al, 2017b, Merelo et al, 2017a, we are going to try and comprehend code repositories by the interactions (events) made in them. Code repositories are one of the main ways software is developed nowadays.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Power laws in code repositories: A skeptical approach

Ortiz,

Merelo-Guervós

2019

Preprint

View full text Add to dashboard Cite

Software development as done using modern methodologies and source control management systems, has been often established as an example of self-organization, with code growing and evolving organically, through activities that do not stem from centralized power, leader or directives. The main challenge in proving these claims is that self organization cannot be detected through direct observation, but through measurements on the system, looking for hints such as the existence of power laws over some features, such as the size of changes over time. The problem we intend to tackle in this paper is to establish a methodology for checking, for a chosen set of repositories we had already measured in the past, if the claims about power laws actually hold from a precise mathematical point of view, since, although shown as pervasive in the software engineering literature (and others), power laws are more elusive than they might seem at first sight. For that reason, in this paper we present a statistically accurate set of tests that will help us decide, from the way repositories are changing, if they are really distributed by a power law, which could indicate us the existence of a state reached via self-organization, or actually, how accurately a power law fits the observed distribution of the size of changes of commits in git repositories of 16 open source repositories. We revisit one of the most representative papers of these observations to reevaluate its results and * This paper has been supported in part by project DeepBio (TIN2017-85727-C4-2-P) 1 compare them with the current status of the repositories analyzed in it, trying to elucidate if there has been any change in the possible presence, or not, of a power law.

show abstract

Section: State Of the Artmentioning

confidence: 99%