Citation Context

... vs. customersrequirements wrt. future maintenance and support. However, it is not clear if PP is well-suitedsto make novices learning from experts [25], or if other team structures are more suitable =-=[5]-=-.sThe paper by Cockburn and Williams enthusiastically summed up the expectations from PPsten years ago [6], cf. also DeMarco’s book on Teamware [7]. But few concrete studies of thesrole of PP in teamw...

Citation Context

...e will have 6 post-release defects of any kind, totally costing 6*34 = 204 p-hs toscorrect. Other industrial studies indicate, that a post-release defect will cost 33-88 p-hs toscorrect, see Humphrey =-=[14]-=-. However, an average, post-release defect-correction cost of suchsmagnitude is simply not credible in the miniature systems from the given PP experiments. Itsmerely illustrates that data from large c...

Citation Context

... Hence, the current PP experiments may prove insufficient tossynthesize a reliable body of evidence for such meta-analysis. Given the Goal QuestionsMetrics principle of lean metrics in empirical work =-=[3]-=-, it should not come as a surprise thatswe cannot come after the completion of a study and demand different or possiblyssupplementary data. However, we can try to agree upon a minimal common metrics f...

Citation Context

... during the last decade [4]. A rather popularsone is Pair Programming (PP), being one of the twelve eXtreme Programming (XP) practices.sThere has similarly been an increase in research efforts (e.g., =-=[26]-=- and [1]) that investigate theseffects of PP, typically by controlled experiments with both professionals and students. Mosts1 This paper was presented at the NIK-2010 conference. For more information...

Citation Context

...ffective teams for industrial-scalesPP. There have been some efforts to perform secondary studies (with informal comparisonssor formal meta-analysis) of primary studies of agile methods, including PP =-=[8, 12]-=-. However,sthe coverage and quality of the published primary studies of agile approaches still needsconsiderable improvement [8]. Hence, the current PP experiments may prove insufficient tossynthesize...

Citation Context

... [25], or if other team structures are more suitable [5].sThe paper by Cockburn and Williams enthusiastically summed up the expectations from PPsten years ago [6], cf. also DeMarco’s book on Teamware =-=[7]-=-. But few concrete studies of thesrole of PP in teamware have been reported from the Agile community. On the effect of humanspersonality on PP performance, there is however a recent study by Hannay et...

Citation Context

...s [9, 13, 16]. This makes research cooperation even harder, if each research groupsgo on “re-inventing the wheel” and even in an inferior way. The overview of controlledsexperiments by Sjøberg et al. =-=[21]-=- lists few replicated experiments, and even worse – not assingle PP experiment. The entire research community has a shared responsibility here.sFurthermore, how can a PhD student get sufficient credit...

Citation Context

...rch methods with varying rigor andsrelevance – from controlled experiments to multi-case studies. There has been a substantialsindustrial uptake of agile approaches and methods during the last decade =-=[4]-=-. A rather popularsone is Pair Programming (PP), being one of the twelve eXtreme Programming (XP) practices.sThere has similarly been an increase in research efforts (e.g., [26] and [1]) that investig...

Citation Context

...he last decade [4]. A rather popularsone is Pair Programming (PP), being one of the twelve eXtreme Programming (XP) practices.sThere has similarly been an increase in research efforts (e.g., [26] and =-=[1]-=-) that investigate theseffects of PP, typically by controlled experiments with both professionals and students. Mosts1 This paper was presented at the NIK-2010 conference. For more information, see //...

Citation Context

...ther 8 studies have data on Correctness and Effort, and 10sstudies have data on Duration and Effort (a rather un-interesting combination).sAnother, less ambitious meta-paper by Stephens and Rosenberg =-=[23]-=- concludes that "pairsprogramming is not uniformly beneficial or effective" ... "you cannot expect faster and bettersand cheaper". The obvious solution is to introduce experience-based cost/benefit mo...

Citation Context

...(non-commenting and non-blank source lines).s61 many such tests were available, e.g., 75 % could mean passing 3 out of 4 tests, or 15 out ofs20.sThe following baseline measures stem from Capers Jones =-=[15]-=- and come from a sample ofs150 large commercial programs with average code size of 2.15 MLOC (megaLOC) and asmedian of 0.28 MLOC. From this sample, Capers Jones predicts that ca. 5 defects persFunctio...

Citation Context

...systems.sTaking a broader perspective, it seems that available (textbook) methods for designing,scarrying out and reporting controlled experiments are not systematically practised by manysresearchers =-=[9, 13, 16]-=-. This makes research cooperation even harder, if each research groupsgo on “re-inventing the wheel” and even in an inferior way. The overview of controlledsexperiments by Sjøberg et al. [21] lists fe...

Citation Context

...systems.sTaking a broader perspective, it seems that available (textbook) methods for designing,scarrying out and reporting controlled experiments are not systematically practised by manysresearchers =-=[9, 13, 16]-=-. This makes research cooperation even harder, if each research groupsgo on “re-inventing the wheel” and even in an inferior way. The overview of controlledsexperiments by Sjøberg et al. [21] lists fe...

Citation Context

... few concrete studies of thesrole of PP in teamware have been reported from the Agile community. On the effect of humanspersonality on PP performance, there is however a recent study by Hannay et al. =-=[11]-=-.s6. ConclusionsBased on this comparative study, we can conclude that there are several kinds of problems insaggregating the findings from PP experiments, mostly due to heterogeneous contextualsvariab...

Citation Context

...tedsto make novices learning from experts [25], or if other team structures are more suitable [5].sThe paper by Cockburn and Williams enthusiastically summed up the expectations from PPsten years ago =-=[6]-=-, cf. also DeMarco’s book on Teamware [7]. But few concrete studies of thesrole of PP in teamware have been reported from the Agile community. On the effect of humanspersonality on PP performance, the...

Citation Context

...systems.sTaking a broader perspective, it seems that available (textbook) methods for designing,scarrying out and reporting controlled experiments are not systematically practised by manysresearchers =-=[9, 13, 16]-=-. This makes research cooperation even harder, if each research groupsgo on “re-inventing the wheel” and even in an inferior way. The overview of controlledsexperiments by Sjøberg et al. [21] lists fe...

Citation Context

...tory in Oslo, and the second one by Williams and her team then at ThesUniversity of Utah in Salt Lake City [26]. In Section 4 we will also discuss the horizontalsdimension, by the attempt of Erdogmus =-=[10]-=- to add commercial defect rates and costs onto thesWilliams’ data. Both PP studies intended to measure the effect of paired vs. solo teams bysvarying several factors, as shown in Table 1 through Table...

Citation Context

...ffective teams for industrial-scalesPP. There have been some efforts to perform secondary studies (with informal comparisonssor formal meta-analysis) of primary studies of agile methods, including PP =-=[8, 12]-=-. However,sthe coverage and quality of the published primary studies of agile approaches still needsconsiderable improvement [8]. Hence, the current PP experiments may prove insufficient tossynthesize...

Citation Context

...ng new technologies or insight in developed software vs. customersrequirements wrt. future maintenance and support. However, it is not clear if PP is well-suitedsto make novices learning from experts =-=[25]-=-, or if other team structures are more suitable [5].sThe paper by Cockburn and Williams enthusiastically summed up the expectations from PPsten years ago [6], cf. also DeMarco’s book on Teamware [7]. ...

Citation Context

...defect. In asrecent Norwegian study we found that a typical correction effort was 8 p-hs in one companyswith a 3 MLOC system, and 11 p-hs in a company with systems of some hundred KLOC, seesLi et al. =-=[18]-=-. We can merely say, that 10 p-hs to correct a post-release coding defect maysserve as an initial quality cost estimate. Note anyhow, that the miniature programs found ins[26] were never intended to h...