Communication in Distributed Agile Development: A Case Study :: Technology, Software Development

Distributed software package development (DSD)1 is becoming acommon practice in modern software industry e.g. 1,where the level of dissemination can consort from teammembers being located in the same city to those ondifferent continents 2. The significance of DSD hasaccelerated because of factors such as improving timetomarketthrough unremitting development across differenttimezones,quick formation of virtual teams and thebenefits of business market advantages. These needs havedriven the software development efforts further towards amultisiteglobally distributed environment. 1Simultaneously, several studies have concluded thatdistributed enterprises are risky e.g. 35.For example, parley and coordination, software quality,schedule overruns and exceeded costs are some of theproblems troubling both singlesiteand distributedsoftware projects. However, the extent of the problem inthe case of DSD seems to be so complex that a thoroughunderstanding of it has not yet been defined. 3, 4 S everalstudies agree that dialogue is a particularlyimportant issue in distributed agile development, e.g. 57.Agile methods rely on volatile requirements that aremanaged through efficient verbal colloquy 8 andthus agile software development methods pose their ownchallenges to the field of DSD.In order to tackle the problems of DSD, severaldifferent techniques have been proposed. Thesetechniques range from utilise different tools, such asinstant messaging 9, videoconferencing 10 andwhiteboard software 5 to a set of more generalrecommendations 5. We conducted two different casestudies with different levels of distribution ranging fromthe customer being in the same city, to one with ageographical distribution of 600 kilometers within thesame country. Therefore, cultural differences were not anissue in these cases. We compared our findings against therecommendations of secular et al. 5 and provide moreinsight on their application based on our empiricalfindings and the existing litera ture. Even though we wereable to evaluate only terzetto recommendations out of theexisting four, our contribution provides valuable insightinto conducting distributed agile projects. Our resultsfurther emphasize the critical role of effectivecommunication, indicating that inefficient and irregularcommunication in conjunction with volatile requirementscan cause severe problems even in very smallscaleagileprojects. However, it seems that effective communicationis not the call. Our cases suggest that having a welldefinedcustomer2 is the key recommendation affecting torecommendations about having a Development Manager5 and using asynchronous communication channels. Asineffective customer collaboration may bed the otherrecommendations redundant, effective customercollaboration seems to be a key factor for successfuldistributed agile development. In addition, wecomplement the existing recommendations by introducingan supererogatory recommendation i.e. enable and supportdirect communic ation between the developers.Unexpectedly, the teams in the second case were notallowed to communicate directly with each other. Tocompensate, a managementledcommunication channelwas established to balance the communication flow,